CN115660796A - Tax fund management method, device, equipment and storage medium for migration risk enterprise - Google Patents

Abstract

The disclosure relates to a tax fund management method, a device, equipment and a storage medium for an emigration risk enterprise, and relates to the technical field of data analysis, wherein the tax fund management method for the emigration risk enterprise comprises the following steps: acquiring the operation data of a target enterprise; performing feature extraction processing on the operation data to obtain features related to enterprise migration behaviors of a target enterprise; inputting the characteristics into a preset enterprise migration risk monitoring model, and detecting migration risks of the target enterprise based on the enterprise migration risk monitoring model. The embodiment of the disclosure can accurately and stably identify and early warn the enterprises with migration risks through the preset enterprise migration risk monitoring model, does not need to depend on human experience, shortens the monitoring time before enterprise migration risk tax, and improves the timeliness and the accuracy of early warning response to loss of key tax sources.

Description

Tax fund management method, device, equipment and storage medium for migration risk enterprise

Technical Field

The present disclosure relates to the field of data analysis technologies, and in particular, to a tax fund management method, device, equipment, and storage medium for an enterprise migration risk.

Background

The tax fund management is the basis of tax collection and management work, and only by strengthening the tax fund management, the tax fund can be guaranteed, and the economy is developed. The main functions of tax source management mainly comprise pre-tax monitoring, collection in tax and post-tax inspection of taxpayer payers. The tax monitoring method has great significance for developing services such as tax source management and tax collection and management and the like in the tax pre-monitoring of enterprise migration risks. The pre-tax monitoring of the enterprise migration risk refers to the detection and identification of potential migration motivations and tendencies of enterprises in the jurisdiction, so that migration risk early warning of key tax sources is realized, and the efficiency of tax collection management and tax payment service work is effectively improved.

At present, a main solution for managing tax sources of enterprises with migratory risks is a monitoring scheme based on an expert evaluation index system, and although the scheme has a certain effect on monitoring before migratory risk taxes of the enterprises, the work of selecting evaluation indexes, setting index weights and risk qualitative thresholds and the like is very dependent on the field experience of tax experts, so that the limitations of poor generalization performance, poor expandability and the like exist, and the monitoring accuracy and efficiency are low.

Disclosure of Invention

In order to solve the technical problem, the present disclosure provides a tax fund management method, device, equipment and storage medium for an emigration risk enterprise.

A first aspect of the embodiments of the present disclosure provides a tax fund management method for an emigration inauguration enterprise, including:

acquiring the operation data of a target enterprise;

performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise;

and inputting the characteristics into a preset enterprise emigration risk monitoring model, and detecting the emigration risk of the target enterprise based on the enterprise emigration risk monitoring model.

A second aspect of the embodiments of the present disclosure provides a tax fund management device for an explanted inauguration enterprise, including:

the acquisition module is used for acquiring the operation data of the target enterprise;

the extraction module is used for carrying out feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise;

and the detection module is used for inputting the characteristics into a preset enterprise emigration risk monitoring model and detecting the emigration risk of the target enterprise based on the enterprise emigration risk monitoring model.

A third aspect of the embodiments of the present disclosure provides a computer device, where the computer device includes a memory and a processor, where the memory stores a computer program, and when the computer program is executed by the processor, the method for managing tax fund of an emigration inauguration enterprise in the first aspect may be implemented.

A fourth aspect of the embodiments of the present disclosure provides a computer-readable storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by a processor, the method for managing tax fund of an emigration inauguration enterprise in the first aspect may be implemented.

Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:

according to the embodiment of the disclosure, the operation data of the target enterprise is obtained; performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise; the method has the advantages that the characteristics are input into the preset enterprise migration risk monitoring model, the migration risk detection is carried out on the target enterprise based on the enterprise migration risk monitoring model, the enterprise with the migration risk can be accurately and stably identified and early warned through the preset enterprise migration risk monitoring model, the artificial experience is not needed, the monitoring time of the enterprise before the migration risk tax is shortened, and the timeliness and the accuracy of early warning response to the loss of the key tax sources are improved.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.

In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present disclosure, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.

Fig. 1 is a flowchart of a tax fund management method for migrating inauguration enterprises according to an embodiment of the present disclosure;

fig. 2 is a flowchart of another tax fund management method for migrating risky enterprises according to an embodiment of the present disclosure;

fig. 3 is a schematic structural diagram of a tax fund management apparatus for migrating risky enterprises according to an embodiment of the present disclosure;

fig. 4 is a schematic structural diagram of a computer device provided in an embodiment of the present disclosure.

Detailed Description

In order that the above objects, features and advantages of the present disclosure may be more clearly understood, aspects of the present disclosure will be further described below. It should be noted that the embodiments and features of the embodiments of the present disclosure may be combined with each other without conflict.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced in other ways than those described herein; it is to be understood that the embodiments disclosed in the specification are only a few embodiments of the present disclosure, and not all embodiments.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.

It is noted that references to "a", "an", and "the" modifications in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.

The tax fund management is the basis of tax collection and management work, and only by strengthening the tax fund management, the tax fund can be guaranteed, and the economy is developed. The main functions of tax source management mainly comprise pre-tax monitoring, collection in tax and post-tax inspection of taxpayer payers. The tax monitoring method has great significance for developing services such as tax source management and tax collection and management and the like in the tax pre-monitoring of enterprise migration risks. The pre-tax monitoring of the enterprise migration risk refers to the detection and identification of potential migration motivations and tendencies of enterprises in the jurisdiction, so that the migration risk early warning of key tax sources is realized, and the efficiency of tax collection management and tax payment service work is effectively improved.

At present, a main solution for managing tax sources of enterprises with migratory risks is a monitoring scheme based on an expert evaluation index system, and although the scheme has a certain effect on monitoring before migratory risk taxes of the enterprises, the work of selecting evaluation indexes, setting index weights and risk qualitative thresholds and the like is very dependent on the field experience of tax experts, so that the limitations of poor generalization performance, poor expandability and the like exist, and the monitoring accuracy and efficiency are low.

Aiming at the defects existing in the monitoring aspect of enterprise migration risk tax in the related art, the embodiment of the disclosure provides a tax fund management method, device, equipment and storage medium for enterprise migration risk, which can accurately and stably identify and early warn the enterprise with migration risk through a preset enterprise migration risk monitoring model, do not need to rely on human experience, shorten the monitoring time of the enterprise before migration risk tax, and improve the timeliness and accuracy of the loss early warning response to key tax fund.

The tax fund management method for migrating risky enterprises provided by the embodiments of the present disclosure may be executed by a computer device, which may be understood as any device with processing capability and computing capability, and the device may include, but is not limited to, a mobile terminal such as a smart phone, a notebook computer, a Personal Digital Assistant (PDA), a tablet computer (PAD), a Portable Multimedia Player (PMP), a vehicle-mounted terminal (e.g., a car navigation terminal), a wearable device, and the like, and a fixed electronic device such as a digital TV, a desktop computer, and the like.

In order to better understand the inventive concept of the embodiments of the present disclosure, the following describes technical solutions of the embodiments of the present disclosure with reference to exemplary embodiments.

Fig. 1 is a flowchart of a tax fund management method for an emigration inauguration enterprise provided by an embodiment of the present disclosure, and as shown in fig. 1, the tax fund management method for an emigration inauguration enterprise provided by this embodiment may include steps 110 to 130:

and step 110, acquiring the operation data of the target enterprise.

The target enterprise in the embodiment of the present disclosure may be understood as a tax payment enterprise registered in any business registration organization.

The operation data in the embodiment of the present disclosure may be understood as operation-related data of an enterprise, and may include tax data and operation-related network data, where the tax data may include data such as a registration address of the enterprise, corporate information, invoicing information, a reporting operation range, financial responsible person information, tax payment reporting information, and shareholder information; the business-related network data may include recruitment information, social networking information, enterprise official networking information, bidding information, and the like of the enterprise.

In the embodiment of the disclosure, the computer device can obtain the tax data of the target enterprise through the tax big data platform, and can obtain the network data related to the operation of the target enterprise through the internet information data platform to obtain the operation data of the target enterprise.

And 120, performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise.

The enterprise migration behavior in the embodiment of the present disclosure may be understood as a behavior in which the enterprise migrates the current operation place from the current jurisdiction to another jurisdiction outside the current jurisdiction, that is, a behavior in which the enterprise changes a tax payment registration mechanism of the jurisdiction corresponding to the enterprise to a tax payment registration mechanism of another jurisdiction outside the jurisdiction.

In the embodiment of the disclosure, the computer device may perform feature extraction processing on the operation data of the target enterprise to obtain features related to the enterprise migration behavior of the target enterprise.

In the embodiment of the disclosure, the characteristics related to the enterprise emigration behavior may include at least one of an office supply purchase abnormal index, a supplier regional distribution abnormal index, a house rental abnormal index, a commercial water consumption amount abnormal index, a commercial electricity consumption amount abnormal index, a commercial trip cost abnormal index, a recruitment abnormal index, an emigration information index, an enterprise major strategic cooperation event frequency, an industry matching index of a jurisdiction where the target enterprise is located, a place of business lease index, and an enterprise legal registration new enterprise index.

The office supply purchasing abnormity index can be understood as the abnormal fluctuation degree of the number of office supplies purchased by an enterprise in a fixed time period, and is used for describing the abnormal degree of the enterprise in terms of office supply purchasing. If the office supply purchase abnormity index of the target enterprise is larger than the first preset threshold value, which indicates that the target enterprise has a large number of office supplies purchased in a fixed time period, the target enterprise can be determined to have the emigration risk.

The supplier geographical distribution abnormality index can be understood as the variation range of the supplier geographical distribution entropy of the enterprise in a fixed time period, and can be used for describing the abnormal fluctuation degree of the supplier geographical distribution of the enterprise and the abnormal degree of the geographical adjustment of the enterprise supply chain. If the supplier region distribution abnormal index of the target enterprise is larger than the second preset threshold, the fact that the suppliers of other jurisdictions outside the jurisdiction where the target enterprise is located in the fixed time period are increased, and the degree of regional adjustment of the suppliers is large is determined, and the target enterprise is determined to have the emigration risk.

The house lease exception index can be used for describing the abnormal behavior of enterprise on the continuous lease of the place of business in a fixed time period.

The commercial water amount abnormal index can depict the abnormal fluctuation of the commercial water invoice amount of the enterprise in a fixed time period. If the abnormal index of the commercial water amount of the target enterprise is larger than the third preset threshold, the abnormal index of the commercial water amount of the target enterprise in a fixed time period is indicated, and the migration risk of the target enterprise can be determined.

The commercial water amount abnormal index can depict the abnormal fluctuation of the commercial power invoice amount of the enterprise in a fixed time period. If the abnormal index of the commercial electricity consumption amount of the target enterprise is larger than the fourth preset threshold, the abnormal index of the commercial electricity consumption amount of the target enterprise in a fixed time period is indicated, and the migration risk of the target enterprise can be determined.

The business trip expense abnormal index can depict the abnormal fluctuation of the expense generated by the enterprise in terms of business trip in a fixed time period. If the business trip expense abnormal index of the target enterprise is larger than the fifth preset threshold, the business trip expense of the target enterprise in a fixed time period is abnormal, and the migration risk of the target enterprise can be determined.

The recruitment abnormal index can depict the abnormal change of regional distribution of the enterprise recruiters. And if the recruitment abnormal index of the target enterprise is larger than the sixth preset threshold, the number of recruiters of other jurisdictions except the jurisdiction where the target enterprise is located is increased in a fixed time period, and the migration risk of the target enterprise can be determined.

The migration information index may be understood as the amount of migration information of an enterprise in the network data within a fixed period of time, and the amount of migration information may be understood as the amount of text information related to enterprise migration in the network data. If the migration information index of the target enterprise is larger than the seventh preset threshold, it is indicated that the number of the migration information of the target enterprise is large, and it can be determined that the target enterprise has a migration risk.

The business major strategic cooperation event frequency rate can be understood as the total event frequency rate of strategic cooperation of the business with other government organizations in the jurisdiction of the business in a fixed time period, and if the business major strategic cooperation event frequency rate of the target business is greater than the eighth preset threshold, it is indicated that the target business has more strategic cooperation events with other government organizations in the jurisdiction of the target business in the fixed time period, and the migration risk of the target business can be determined.

The industry matching index can be understood as the matching degree of the industry structure in the district where the enterprise is located and the main business of the target enterprise in a fixed time period. If the industry matching index of the target enterprise is larger than the ninth preset threshold, the matching degree of the industry structure in the district where the target enterprise is located and the main business of the target enterprise in a fixed time period is poor, and the migration risk of the target enterprise can be determined.

The business floor rental index can describe new rental behavior of the enterprise registered business floor in a fixed time period.

The enterprise jurisdictional registration of the new enterprise index can identify new enterprise registration behaviors of the enterprise jurisdictions other than the jurisdiction where the target enterprise is located within a fixed time period.

In some embodiments, the business data for the target enterprise may include data on the amount of office supplies purchased by the target enterprise. Performing feature extraction processing on the operation data to obtain features related to enterprise migration behaviors of the target enterprise, wherein the features may include S11-S15:

s11, determining a first purchase amount of the target enterprise for purchasing office supplies every day in preset days based on the amount data of the office supplies purchased by the target enterprise, and calculating an average purchase amount corresponding to the first purchase amount in the preset days.

In the embodiment of the disclosure, the computer device may acquire the amount data of office supplies purchased by the target enterprise from the invoice data of office supplies purchased by the target enterprise, then determine the first purchase amount of office supplies purchased by the target enterprise each day within the preset number of days, and determine the ratio of the sum of the purchase amounts of office supplies purchased by the target enterprise within the preset number of days to the preset number of days as the average purchase amount of office supplies purchased by the target enterprise within the preset number of days. The preset number of days may be set as needed, for example, 365 days, and is not limited herein.

And S12, calculating the standard deviation of the purchasing amount of the office supplies purchased by the target enterprise within the preset days based on the first purchasing amount and the average purchasing amount.

In the embodiment of the disclosure, the computer device may calculate the standard deviation of the purchase amount of the target enterprise for purchasing the office supplies within the preset number of days according to the purchase amount and the average purchase amount within the preset number of days.

S13, calculating a first difference value between a second purchasing amount of the target enterprise for purchasing office supplies every day and the average purchasing amount within the target days, and calculating the square of a first ratio of the first difference value to the standard deviation of the purchasing amount.

In the embodiment of the disclosure, the computer device may obtain a second purchase amount of the target enterprise for purchasing the office supplies each day within the target number of days, calculate a first difference between the second purchase amount of the target enterprise for purchasing the office supplies each day within the target number of days and the average purchase amount, calculate a first ratio of the first difference to a standard deviation of the purchase amount, and calculate a square of the first ratio. The target number of days may be set according to actual needs, and is not particularly limited herein.

And S14, summing the squares of the first specific values corresponding to each day in the target days to obtain a first square sum of the target days, and determining the first square sum as an office supply purchase abnormity index of the target enterprise in the target days.

In the embodiment of the disclosure, the computer device may sum the squares of the first ratio corresponding to each day in the target number of days to obtain a first square sum of the target number of days, and determine the first square sum as an index of office supply purchase anomaly of the target enterprise in the target number of days.

And S15, determining office supply purchase abnormity indexes of the target enterprise in the target days as characteristics of the target enterprise related to enterprise migration behaviors.

In the embodiment of the disclosure, the computer device may determine the office supply procurement anomaly index of the target enterprise within the target days as a characteristic of the target enterprise related to the enterprise migration behavior.

For example, the formula for calculating the office supply procurement anomaly index of the target enterprise can be expressed as:

wherein the content of the first and second substances,

indicating the index of abnormal office supply procurement of the target enterprise within the target number of days,

the second purchase amount of the office supplies purchased by the target enterprise on the ith day representing the target number of days,

representing the average purchasing amount corresponding to the first purchasing amount of the target enterprise within the preset number of days,

and K represents the target number of days.

In other embodiments, the business data for the target enterprise may include the total invoiced amounts of suppliers received by the target enterprise on a daily basis and the invoiced amounts of suppliers for various jurisdictions received by the target enterprise on a daily basis, as well as the supplier data for the target enterprise. Performing feature extraction processing on the operation data to obtain features of the target enterprise related to enterprise migration behaviors, which may include S21-S26:

s21, calculating a second ratio of the invoicing amount of the target enterprise in each district per day to the total invoicing amount per day within the preset number of days based on the total invoicing amount of the supplier received by the target enterprise per day and the invoicing amount of the supplier in each district received by the target enterprise per day.

In the embodiment of the disclosure, the computer device may determine the invoicing amount of the provider in each jurisdiction and the total invoicing amount of the provider in each jurisdiction, which are received by the target enterprise every day, within a preset number of days, and calculate a second ratio of the invoicing amount of the provider in each jurisdiction and the total invoicing amount of the provider in each day, which are received by the target enterprise within the preset number of days.

And S22, calculating the supplier region distribution entropy corresponding to the target enterprise every day in the preset days and the first average distribution entropy corresponding to the supplier region distribution entropy in the preset days based on the second ratio.

In the embodiment of the disclosure, the supplier geographical distribution entropy may be understood as a variation size of the supplier geographical distribution, if the enterprise has no migration risk, the geographical distribution of the enterprise supplier is stable, the supplier geographical distribution entropy of the enterprise is smaller, and if the enterprise has the migration risk, suppliers outside the jurisdiction of the enterprise may increase, the supplier geographical distribution entropy of the enterprise may increase.

In the embodiment of the disclosure, the computer device may calculate, based on the second ratio, a provider regional distribution entropy corresponding to the target enterprise every day within a preset number of days, and calculate a ratio of a sum of the provider regional distribution entropies within the preset number of days to the preset number of days, so as to obtain the first average distribution entropy.

S23, calculating a first distribution entropy standard deviation of the supplier region distribution entropy in preset days based on the supplier region distribution entropy and the first average distribution entropy.

In the embodiment of the disclosure, the computer device may calculate a first distribution entropy standard deviation of the supplier geographical distribution entropy within a preset number of days based on the supplier geographical distribution entropy and the first average distribution entropy.

S24, calculating an absolute value of a second difference value between the provider region distribution entropy and the first average distribution entropy of the target enterprise in each day in the target days, and calculating a third ratio of the absolute value of the second difference value to the standard deviation of the first distribution entropy.

In an embodiment of the disclosure, the computer device may calculate an absolute value of a second difference between the provider geographical distribution entropy and the first average distribution entropy for each day of the target enterprise within the target number of days, and calculate a third ratio between the absolute value of the second difference and the standard deviation of the first distribution entropy.

And S25, summing the third ratio corresponding to each day in the target days to obtain a third ratio sum in the target days, and determining the third ratio sum as the provider regional distribution abnormal index of the target enterprise in the target days.

In this disclosure, the computer device may sum the third ratio corresponding to each day in the target number of days to obtain a sum of the third ratios in the target number of days, and determine the sum of the third ratios as an abnormal index of the distribution of the provider regions of the target enterprise in the target number of days.

And S26, determining the provider regional distribution abnormal index of the target enterprise in the target days as the characteristic related to the enterprise migration behavior of the target enterprise.

In the embodiment of the disclosure, the computer device may determine the provider regional distribution abnormal index of the target enterprise within the target number of days as a characteristic of the target enterprise related to the enterprise emigration behavior.

For example, the calculation formula of the provider regional distribution abnormality index of the target enterprise in the target days can be expressed as:

wherein, the first and the second end of the pipe are connected with each other,

indicating the provider regional distribution anomaly index of the target enterprise within the target number of days,

the supplier geographic distribution entropy of the target business on day i representing the target number of days,

representing a first average distribution entropy corresponding to the supplier region distribution entropy of the target enterprise within the preset number of days,

and K represents the target number of days. Wherein the content of the first and second substances,

a second ratio of the supplier's invoiced amount received by the target business on day i in the jth jurisdiction to the total invoiced amount received by the target business on day i, representing the target number of days.

In other embodiments, the business data for the target business may include house rental invoice data for the target business. The method comprises the steps of performing feature extraction processing on operation data to obtain features related to enterprise migration behaviors of a target enterprise, determining whether the target enterprise has an operation place continuous lease postponing behavior within a target number of days or not by computer equipment based on house lease invoice data of the target enterprise, determining the house lease abnormal index of the target enterprise within the target number of days as 1 by the computer equipment if the target enterprise has the operation place continuous lease postponing behavior within the target number of days, determining the house lease abnormal index of the target enterprise within the target number of days as 0 by the computer equipment if the target enterprise does not have the operation place continuous lease postponing behavior within the target number of days, and determining the house lease abnormal index of the target enterprise within the target number of days as the features related to the enterprise migration behaviors of the target enterprise.

In other embodiments, the business data for the target enterprise may include business water usage data for the target enterprise. Performing feature extraction processing on the operation data to obtain features related to enterprise migration behaviors of the target enterprise, wherein the steps include S31-S35:

s31, determining the first commercial water amount of the target enterprise in each day in the preset days based on the commercial water amount data of the target enterprise, and calculating the average water amount corresponding to the first commercial water amount in the preset days.

In the embodiment of the disclosure, the computer device may determine a first commercial water amount of the target enterprise per day within a preset number of days based on the commercial water amount data of the target enterprise, and calculate a ratio of a sum of the first commercial water amounts within the preset number of days to obtain an average water amount.

And S32, calculating the standard deviation of the water consumption amount of the commercial water for the target enterprise in the preset days based on the first commercial water consumption amount and the average water consumption amount.

In the embodiment of the disclosure, the computer device may calculate the standard deviation of the water usage amount of the commercial water for the target enterprise within the preset number of days based on the first commercial water usage amount and the average water usage amount of the target enterprise per day within the preset number of days.

And S33, calculating the square of a fourth ratio of the third difference to the standard deviation of the water usage amount based on the third difference between the second commercial water usage amount and the average water usage amount of the target enterprise in each day in the target days.

In embodiments of the present disclosure, the computer device may calculate a third difference between the second commercial use amount and the average use amount for each day of the target enterprise over the target number of days, then calculate a fourth ratio of the third difference to the standard deviation of the use amount, and calculate a square of the fourth ratio.

And S34, summing the squares of the fourth specific values corresponding to each day in the target days to obtain a second square sum in the target days, and determining the second square sum as the commercial water use amount abnormal index of the target enterprise in the target days.

In the embodiment of the disclosure, the computer device may sum the squares of the fourth ratio corresponding to each day in the target number of days to obtain a second sum of squares in the target number of days, and determine the second sum of squares as the business water amount abnormality index of the target enterprise in the target number of days.

And S35, determining the business water amount abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise emigration behavior.

In the disclosed embodiment, the computer device may determine the business water amount anomaly index of the target enterprise within the target number of days as a characteristic of the target enterprise related to the enterprise migration behavior.

For example, the calculation formula of the business water amount abnormality index of the target enterprise in the target days can be expressed as:

wherein the content of the first and second substances,

a business water amount anomaly index representing the target business over the target number of days,

a second commercial water amount for the target business on day i representing the target number of days,

the average water consumption corresponding to the first commercial water consumption of the target enterprise in the preset days is represented,

the standard deviation of the water consumption amount of the commercial water in the preset days of the target enterprise is shown, and K represents the target days.

In other embodiments, the business data for the target enterprise may include business electricity amount data for the target enterprise. Performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise, wherein the steps include S41-S45:

s41, determining the first commercial electricity consumption amount of the target enterprise in each day in the preset days based on the commercial electricity consumption amount data of the target enterprise, and calculating the average electricity consumption amount corresponding to the first commercial electricity consumption amount in the preset days.

In the embodiment of the disclosure, the computer device may determine, based on the commercial power amount data of the target enterprise, a first commercial power amount of the target enterprise every day in a preset number of days, and calculate a ratio of a sum of the first commercial power amount in the preset number of days to the preset number of days, to obtain an average power amount.

And S42, calculating the standard deviation of the electricity consumption amount of the commercial electricity consumption of the target enterprise in the preset number of days based on the first commercial electricity consumption amount and the average electricity consumption amount.

In the embodiment of the disclosure, the computer device may calculate the standard deviation of the electricity consumption amount of the commercial electricity of the target enterprise within the preset number of days based on the first commercial electricity consumption amount and the average electricity consumption amount.

And S43, calculating the square of a fifth ratio of the fourth difference to the standard deviation of the electricity consumption amount based on the fourth difference between the second commercial electricity consumption amount and the average electricity consumption amount of the target enterprise in each day in the target days.

In an embodiment of the present disclosure, the computer device may calculate a fifth ratio of the fourth difference to the standard deviation of the electricity usage amount based on a fourth difference between a second commercial electricity usage amount and the average electricity usage amount of the target enterprise per day for the target number of days, and calculate a square of the fifth ratio.

And S44, summing the squares of the fifth specific values corresponding to each day in the target days to obtain a third square sum in the target days, and determining the third square sum as the commercial electricity sum abnormal index of the target enterprise in the target days.

In the embodiment of the disclosure, the computer device may sum squares of the fifth ratio corresponding to each day in the target number of days to obtain a third sum of squares in the target number of days, and determine the third sum of squares as an abnormal index of the commercial power amount of the target enterprise in the target number of days.

And S45, determining the abnormal index of the commercial electricity consumption amount of the target enterprise in the target days as the characteristic related to the migration behavior of the target enterprise and the enterprise.

In the embodiment of the disclosure, the computer device may determine the abnormal index of the commercial electricity consumption amount of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

For example, the calculation formula of the business water amount abnormality index of the target enterprise in the target days can be expressed as:

wherein the content of the first and second substances,

indicating the abnormal index of the commercial power amount of the target enterprise in the target days,

a second amount of commercial power for the target business on day i representing the target number of days,

representing the average electricity consumption corresponding to the first commercial electricity consumption of the target enterprise in the preset days,

and K represents the target number of days.

In other embodiments, the business data of the target enterprise may include business trip fee data of the target enterprise. Performing feature extraction processing on the operation data to obtain features of the target enterprise related to enterprise migration behaviors, which may include S51-S55:

s51, determining first business trip cost of the target enterprise in preset days based on business trip cost data of the target enterprise, and calculating average business trip cost corresponding to the first business trip cost in the preset days.

In the embodiment of the disclosure, the computer device may determine, based on business trip cost data of the target enterprise, first business trip costs of the target enterprise in preset days per day, and calculate a ratio of a sum of the first business trip costs in the preset days to the preset days, so as to obtain an average business trip cost.

And S52, calculating the standard deviation of the business trip cost of the target enterprise on the business trip within the preset number of days based on the first business trip cost and the average business trip cost.

In the embodiment of the disclosure, the computer device may calculate a standard deviation of business trip costs of the target enterprise for business trips within a preset number of days based on the first business trip cost and the average business trip cost.

And S53, calculating the square of a sixth ratio of the fifth difference to the standard deviation of the business trip expenses based on the fifth difference between the second business trip expenses of the target enterprise in each day and the average business trip expenses in the target days.

In the embodiment of the present disclosure, the computer device may calculate a sixth ratio of the fifth difference to the standard deviation of the business trip costs based on a fifth difference between the second business trip costs of the target enterprise per day and the average business trip costs in the target number of days, and calculate a square of the sixth ratio.

And S54, summing the squares of the sixth specific values corresponding to each day in the target days to obtain a fourth square sum in the target days, and determining the fourth square sum as the business trip cost abnormal index of the target enterprise in the target days.

In this disclosure, the computer device may sum the squares of the sixth ratio corresponding to each day in the target number of days to obtain a fourth sum of squares in the target number of days, and determine the fourth sum of squares as the business trip cost anomaly index of the target enterprise in the target number of days.

And S55, determining the business trip cost abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

In the embodiment of the disclosure, the computer device may determine the business trip cost abnormal index of the target enterprise within the target number of days as a characteristic of the target enterprise related to the enterprise migration behavior.

For example, the calculation formula of the business trip cost anomaly index of the target enterprise in the target days can be expressed as follows:

wherein the content of the first and second substances,

representing the business trip cost abnormal index of the target enterprise in the target days,

a second business trip fee for the target business on day i representing the target number of days,

representing the average business trip cost corresponding to the first business trip cost of the target enterprise within the preset days,

and K represents the standard deviation of the business trip cost of the business trip of the target enterprise within the preset number of days, and K represents the target number of days.

In other embodiments, the business data of the target enterprise may include a total number of recruiters for target recruitment every day for the target enterprise and a number of recruiters for target recruitment in each jurisdiction every day for the target enterprise. Performing feature extraction processing on the business data to obtain features of the target enterprise related to enterprise migration behaviors, which may include S61-S66:

s61, calculating a seventh ratio of the number of the recruiting persons of the target enterprise in each jurisdiction every day to the number of the total of the recruiting persons of the target enterprise in each jurisdiction every day based on the total number of the recruiting persons of the target enterprise in each jurisdiction every day and the number of the recruiting persons of the target enterprise in each jurisdiction every day.

In the embodiment of the disclosure, the computer device may calculate a seventh ratio of the number of recruiting persons for target recruitment of the target enterprise in each jurisdiction to the number of total recruiting persons for target recruitment of the target enterprise in each jurisdiction every day within a preset number of days based on the total recruiting persons for target recruitment of the target enterprise every day and the number of recruiting persons for target recruitment of the target enterprise in each jurisdiction every day.

And S62, calculating a regional distribution entropy of the recruiters for the target recruitment every day of the target enterprise within the preset days and a second average distribution entropy corresponding to the regional distribution entropy of the recruiters within the preset days based on the seventh ratio.

In the embodiment of the disclosure, the geographical distribution entropy of the recruiters can be understood as the variation size of the geographical distribution of the recruiters, if the enterprise has no migration risk, the geographical distribution of the recruiters of the enterprise is stable, the geographical distribution entropy of the recruiters of the enterprise is smaller, and if the enterprise has the migration risk, the geographical distribution entropy of the recruiters outside the jurisdiction of the enterprise is increased, and the geographical distribution entropy of the recruiters of the enterprise is increased.

In the embodiment of the disclosure, the computer device may calculate, based on the seventh ratio, a regional distribution entropy of the recruiters for target recruitment every day of the target enterprise within a preset number of days, and calculate a ratio of a sum of the regional distribution entropies of the recruiters within the preset number of days to the preset number of days, so as to obtain a second average distribution entropy.

And S63, calculating a second distribution entropy standard deviation of the regional distribution entropy of the recruiters in the preset days based on the regional distribution entropy of the recruiters and the second average distribution entropy.

In the embodiment of the disclosure, the computer device may calculate a second distribution entropy standard deviation of the regional distribution entropy of the recruiter within the preset number of days based on the regional distribution entropy of the recruiter and the second average distribution entropy.

And S64, calculating an absolute value of a sixth difference value between the regional distribution entropy of the recruiters and the second average distribution entropy of the target enterprise in each day within the target days, and calculating an eighth ratio of the sixth difference value to the standard deviation of the second distribution entropy.

In the embodiment of the disclosure, the computer device may calculate an absolute value of a sixth difference between the regional distribution entropy of the recruiters and the second average distribution entropy of the target enterprise each day within the target number of days, and calculate an eighth ratio between the sixth difference and the standard deviation of the second distribution entropy.

And S65, summing the eighth ratio corresponding to each day in the target days to obtain an eighth ratio sum in the target days, and determining the eighth ratio sum as the recruitment abnormal index of the target enterprise in the target days.

In the embodiment of the disclosure, the computer device may sum the eighth ratio corresponding to each day in the target number of days to obtain an eighth ratio sum in the target number of days, and determine the eighth ratio sum as a recruitment abnormality index of the target enterprise in the target number of days.

S66, determining the recruitment abnormal index of the target enterprise in the target days as the characteristic related to the migration behavior of the target enterprise and the enterprise.

In the embodiment of the disclosure, the computer device may determine the recruitment abnormality index of the target enterprise within the target number of days as a characteristic of the target enterprise related to the enterprise migration behavior.

For example, the calculation formula of the recruitment anomaly index of the target enterprise in the target days can be expressed as:

wherein, the first and the second end of the pipe are connected with each other,

indicating the recruitment anomaly index of the target enterprise within the target number of days,

regional distribution entropy of recruiters of target enterprises on the ith day representing the target days,

representing a second average distribution entropy corresponding to the regional distribution entropy of the recruiters of the target enterprise within the preset days,

and K represents the target number of days. Wherein the content of the first and second substances,

day i target Business representing target day number

And the seventh ratio of the recruiter number of the target recruitments of each jurisdiction to the total recruiter number of the target recruitments of the target enterprises on the ith day.

In other embodiments, the business data for the target enterprise may include the amount of enterprise migrations information for the target enterprise. The method comprises the steps of extracting and processing features of business data to obtain features of a target enterprise related to enterprise migration behaviors, summing the number of target migration information within target days by computer equipment based on the number of enterprise migration information of the target enterprise to obtain the total number of migration information within the target days, determining the total number of migration information as the index of migration information of the target enterprise within the target days, and determining the index of migration information of the target enterprise within the target days as the features of the target enterprise related to the enterprise migration behaviors.

In other embodiments, the business data of the target enterprise may include event frequency of strategic cooperation between the target enterprise and government organizations outside the jurisdiction of the target enterprise, which may be understood as the number of strategic cooperation events between the enterprise official network issues and government organizations outside the jurisdiction of the enterprise. The method comprises the steps of extracting characteristics of business data to obtain characteristics related to migration behaviors of a target enterprise and the target enterprise, obtaining event frequency of strategic cooperation between the target enterprise and other government organizations outside the jurisdiction where the target enterprise is located by computer equipment from an enterprise official network of the target enterprise, summing the event frequency of the strategic cooperation between the target enterprise and the other government organizations within the target day to obtain total event frequency within the target day, determining the total event frequency as the enterprise major strategic cooperation event frequency of the target enterprise within the target day, and determining the enterprise major strategic cooperation event frequency of the target enterprise within the target day as the characteristics related to the migration behaviors of the target enterprise and the target enterprise.

In other embodiments, the business data for the target enterprise may include upstream and downstream enterprise data for the target enterprise. Performing feature extraction processing on the operation data to obtain features of the target enterprise related to enterprise emigration behaviors, which may include S71-S74:

s71, based on the upstream and downstream enterprise data of the target enterprise, calculating a first proportion of the number of the foreign enterprises in the upstream enterprise of the target enterprise and a second proportion of the number of the foreign enterprises in the downstream enterprise of the target enterprise in the number of the downstream enterprise.

In the embodiment of the present disclosure, an upstream enterprise may be understood as a supplier of an enterprise, a downstream enterprise may be understood as a client of the enterprise, and an alien enterprise may be understood as another enterprise outside the jurisdiction where the enterprise is located.

In embodiments of the present disclosure, a computer device may calculate a first percentage of the number of foreign businesses in an upstream business of a target business and a second percentage of the number of foreign businesses in a downstream business of the target business over a target number of days in a number of upstream businesses based on upstream and downstream business data for the target business.

And S72, calculating a harmonic mean of the first ratio and the second ratio.

In embodiments of the present disclosure, a computer device may calculate a harmonic mean of a first and second ratio.

And S73, determining the harmonic mean as the industry matching index of the district where the target enterprise is located in the target days.

In the embodiment of the disclosure, the computer device may determine the harmonic mean of the first proportion and the second proportion as an industry matching index of the jurisdiction where the target enterprise is located within the target days.

And S74, determining the industry matching index of the target enterprise in the district in which the target enterprise is located in the target days as the characteristic related to the migration behavior of the target enterprise and the enterprise.

In the embodiment of the disclosure, the computer device may determine the industry matching index of the target enterprise in the jurisdiction of the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

For example, the calculation formula of the industry matching index of the target enterprise in the jurisdiction of the target day may be represented as follows:

wherein, the first and the second end of the pipe are connected with each other,

the business matching index of the district where the target enterprise is located in the target day number is represented, Q represents a first ratio of the number of foreign enterprises in the upstream enterprises of the target enterprise in the target day number, and R represents a second ratio of the number of foreign enterprises in the downstream enterprises of the target enterprise.

In other embodiments, the business data of the target enterprise may include rental information for the business premises of the target enterprise. The method comprises the steps of performing feature extraction processing on operation data to obtain features related to enterprise migration behaviors of a target enterprise, determining whether the renting information of the operation place of the target enterprise exists in target days based on the renting information of the operation place of the target enterprise, determining the operation place renting index of the target enterprise in the target days based on whether the renting information of the operation place of the target enterprise exists in the target days, wherein if the renting information of the operation place of the target enterprise exists in the target days, the operation place renting index of the target enterprise in the target days can be 1, if the renting information of the operation place of the target enterprise does not exist in the target days, the operation place renting index of the target enterprise in the target days can be 0, and determining the operation place renting index of the target enterprise in the target days as the features related to the enterprise migration behaviors.

In other embodiments, the business data of the target enterprise may include registered new enterprise behavior data of enterprise jurisdictions of the target enterprise other than the jurisdiction in which the target enterprise is located. Performing feature extraction processing on the operation data to obtain features related to enterprise migration behaviors of a target enterprise, and determining whether the enterprise legal person of the target enterprise has a behavior of registering a new enterprise in other jurisdictions outside the jurisdiction of the target enterprise or not based on the registered new enterprise behavior data of the enterprise legal person of the target enterprise in other jurisdictions outside the jurisdiction of the target enterprise within the target days by using the computer equipment; and determining the index of the enterprise legal person of the target enterprise registering the new enterprise in the target days based on whether the enterprise legal person of the target enterprise has the action of registering the new enterprise in other jurisdictions except the jurisdiction of the target enterprise in the target days. If the enterprise legal person of the target enterprise has the behavior of registering the new enterprise in other jurisdictions except the jurisdiction where the target enterprise is located within the target number of days, the index of registering the new enterprise by the enterprise legal person of the target enterprise within the target number of days can be determined to be 1, if the behavior of registering the new enterprise does not exist in other jurisdictions except the jurisdiction where the target enterprise is located within the target number of days, the index of registering the new enterprise by the enterprise legal person of the target enterprise within the target number of days can be determined to be 0, and the index of registering the new enterprise by the enterprise legal person of the target enterprise within the target number of days can be determined to be the characteristic related to the enterprise migratory behavior of the target enterprise.

And step 130, inputting the characteristics into a preset enterprise migration risk monitoring model, and detecting the migration risk of the target enterprise based on the enterprise migration risk monitoring model.

In the embodiment of the disclosure, the computer device may build an enterprise emigration risk monitoring model through model training in advance.

In some embodiments of the present disclosure, after obtaining the characteristics of the target enterprise related to the enterprise migration behavior, the computer device needs to encode the characteristics to obtain a characteristic vector corresponding to the characteristics, then input the characteristic vector into a preset enterprise migration risk monitoring model, and perform migration risk detection on the target enterprise based on the enterprise migration risk monitoring model to obtain a result of whether the target enterprise has a migration risk.

In other embodiments of the present disclosure, after obtaining the characteristics of the target enterprise related to the enterprise migration behavior, the computer device may input the characteristics into a preset enterprise migration risk monitoring model, and perform migration risk detection on the target enterprise based on the enterprise migration risk monitoring model to obtain a result of whether the target enterprise has a migration risk.

In the embodiment of the present disclosure, the enterprise emigration risk monitoring model may include a model constructed based on a gradient lifting tree, the gradient lifting tree may be understood as a classifier that integrates multiple decision tree models based on a lifting algorithm (boosting) integration strategy, the gradient lifting tree may integrate classification and aggregation results of all decision trees, and assign and output aggregation results output by all base learners as final class labels, thereby ensuring higher classification accuracy and ensuring robustness of a classification process.

According to the embodiment of the disclosure, the operation data of the target enterprise is obtained; performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise; the method has the advantages that the characteristics are input into the preset enterprise migration risk monitoring model, the migration risk detection is carried out on the target enterprise based on the enterprise migration risk monitoring model, the enterprise with the migration risk can be accurately and stably identified and early warned through the preset enterprise migration risk monitoring model, the artificial experience is not needed, the monitoring time of the enterprise before the migration risk tax is shortened, and the timeliness and the accuracy of early warning response to the loss of the key tax sources are improved.

Fig. 2 is a flowchart of a tax fund management method for an migrating venture enterprise provided in an embodiment of the present disclosure, and as shown in fig. 2, the tax fund management method for an migrating venture enterprise provided in this embodiment may include steps 210 to 260:

step 210, obtaining the operation data of a plurality of enterprises and the label data of a preset number of enterprises in the plurality of enterprises, and constructing an original data set based on the operation data of the preset number of enterprises, wherein the plurality of enterprises and the preset number of enterprises comprise migrated enterprises and non-migrated enterprises.

In the embodiment of the disclosure, the computer device may obtain the operation data of the plurality of enterprises and the tag data of a preset number of enterprises in the plurality of enterprises, and construct the original data set based on the operation data of the preset number of enterprises, where the plurality of enterprises and the preset number of enterprises include migrated enterprises and non-migrated enterprises. The preset number may be set as needed, and is not particularly limited herein.

And step 220, training the model by adopting an active learning strategy based on the original data set to obtain a preset enterprise emigration risk monitoring model.

In the embodiment of the disclosure, the active learning strategy can be understood as a model training method for hierarchically labeling data in an iterative manner in a scene with higher sample labeling cost, so that more labeled resources are poured into the sample with higher modeling difficulty, the use efficiency of labeled resources can be improved, the effectiveness of modeling in a small sample scene is improved, and the labeling cost is reduced.

In the embodiment of the disclosure, the computer device may train the model by using an active learning strategy based on the original data set to obtain a preset enterprise emigration risk monitoring model.

In some embodiments, training the model by using an active learning strategy based on the original data set to obtain a preset enterprise migration risk monitoring model may include steps 2201-2207:

step 2201, inputting a first data set composed of the operation data and the label data of a preset number of enterprises into a model, and performing model training based on the first data set to obtain a first training model.

In the embodiment of the disclosure, the computer device may input a first data set composed of the operation data and the tag data of a preset number of enterprises into the model, and perform model training based on the first data set to obtain a first training model.

Step 2202, inputting the residual data sets except the first data set in the original data set into a first training model, and predicting the enterprise samples in the residual data sets based on the first training model to obtain the prediction results of all the enterprise samples in the residual data sets.

In the embodiment of the disclosure, the computer device may input the remaining data sets except the first data set in the original data set into the first training model, and predict the enterprise samples in the remaining data sets based on the first training model to obtain the prediction results of the enterprise samples in the remaining data sets. For example, according to the probability that each enterprise sample is migrated enterprise and non-migrated enterprise predicted by the first training model, the information entropy corresponding to each enterprise sample is determined, if the difference between the probability of migrating enterprise and the probability of non-migrated enterprise is smaller, the prediction result of the model is inaccurate, the information entropy of the enterprise sample is larger, and if the difference between the probability of migrating enterprise and the probability of non-migrated enterprise is larger, the prediction result of the model is more accurate, the information entropy of the enterprise sample is smaller. The information entropy can be understood as a parameter describing whether the distribution is concentrated or not.

And 2203, screening the enterprise samples with wrong prediction from the residual data set as the enterprise samples to be labeled based on the prediction results of the enterprise samples in the residual data set.

In the embodiment of the disclosure, the computer device may screen, from the remaining data set, the enterprise sample with the incorrect prediction as the enterprise sample to be labeled, based on the prediction result of each enterprise sample in the remaining data set.

In some embodiments, the information entropies of the enterprise samples in the remaining data sets may be sorted in descending order to obtain an information entropy sequence, and the enterprise samples corresponding to the information entropies of the front target number in the information entropy sequence are selected as the enterprise samples to be labeled.

In other embodiments, the enterprise samples with the information entropy greater than the preset threshold may be selected as the enterprise samples to be labeled according to the information entropy of each enterprise sample in the remaining data set.

Step 2204, based on manual labeling operation, determining label data of the enterprise sample to be labeled to obtain a labeled enterprise sample.

In the embodiment of the disclosure, the marking operation can be performed on the enterprise sample to be marked manually, and the computer device can determine the label data of the enterprise sample to be marked based on the manual marking operation to obtain the marked enterprise sample.

Step 2205, adding the marked enterprise sample into the first data set to obtain a second data set.

In an embodiment of the present disclosure, the computer device may add the annotated enterprise sample to the first data set to obtain a second data set.

Step 2206, training the first training model based on the second data set to obtain a second training model.

In an embodiment of the present disclosure, the computer device may train the first training model based on the second data set to obtain a second training model.

And 2207, sequentially performing the steps of prediction, screening, determination, addition and training on the enterprise samples in the residual data set to obtain a trained enterprise migration risk monitoring model, and determining the trained enterprise migration risk monitoring model as a preset enterprise migration risk monitoring model.

And step 230, acquiring the operation data of the target enterprise.

The steps of the embodiment of the present disclosure may refer to the content of step 110, which is not described herein again.

And 240, preprocessing the operation data, wherein the preprocessing mode at least comprises one of missing value processing, abnormal value processing, dimension processing, duplication removing processing and noise processing.

In the embodiment of the disclosure, after obtaining the business data of the target enterprise, the computer device may preprocess the business data to obtain the preprocessed business data. The preprocessing mode at least comprises one of missing value processing, abnormal value processing, dimension processing, de-duplication processing and noise processing.

The missing value in the embodiment of the present disclosure may be understood as data missing in the business data due to some reason, and it may be understood that data having the missing value is incomplete data. The reasons for generating the missing value mainly include a mechanical reason and an artificial reason, wherein the mechanical reason is data missing caused by failure of data collection or storage caused by the machine itself, such as failure of data storage, memory damage, failure of the machine, and the like, which result in failure to collect certain data. The human cause is data loss due to human subjective error, history limitation or intentional concealment, for example, data entry personnel miss data entry. In some embodiments, when the ratio of the missing value in the operation data is greater than or equal to a first preset ratio threshold, the operation data may be considered to reduce the detection accuracy, and the operation data may be subjected to a rejection process; in some embodiments, when the ratio of the missing value in the operation data is smaller than the first preset ratio threshold, the missing value in the operation data may be subjected to data supplementation so as to complete the operation data.

Abnormal values in the embodiments of the present disclosure may be understood as erroneous data, and the cause of occurrence may include data logging personnel mistyping the data. In some embodiments, when the proportion of the abnormal value in the tax data is greater than or equal to a second preset proportion threshold, the business data may be considered to reduce the detection accuracy, and the business data may be subjected to a rejection process; when the proportion of the abnormal value in the operation data is smaller than the second preset proportion threshold value, the operation data can be considered not to influence the detection accuracy, and the operation data can be reserved.

Dimensions in the embodiments of the present disclosure may be understood as measurable physical properties inherent to physical quantities of data, each physical quantity of data having only one dimension. Dimension processing may be understood as unifying the dimensions of the respective data into one and the same dimension, for example, unifying the unit of money "dollars" into "RMB", facilitating subsequent calculation.

The deduplication processing in the embodiments of the present disclosure may be understood as removing data that appears repeatedly.

The noise processing in the embodiments of the present disclosure may be understood as removing data that is not related to the business data.

According to the embodiment of the invention, the operational data is preprocessed, so that the operational data can be standardized, the influence of data factors on the reasoning performance of the enterprise migration risk monitoring model is reduced, and the detection accuracy of the enterprise migration risk monitoring model is improved.

And 250, performing feature extraction processing on the operation data to obtain features related to the enterprise migration behavior of the target enterprise.

And 260, inputting the characteristics into a preset enterprise migration risk monitoring model, and detecting the migration risk of the target enterprise based on the enterprise migration risk monitoring model.

The contents of steps 250-260 in the embodiment of the present disclosure may refer to steps 120-130 described above, and are not described herein again.

From this, can train enterprise migration risk monitoring model through initiative learning strategy, reduce the data bulk of mark, improve marking efficiency, reduce the mark cost, through carrying out the preliminary treatment to the operation data, reduce the influence of data factor to enterprise migration risk monitoring model reasoning performance, improve the accuracy that enterprise migration risk monitoring model detected, can carry out accurate identification and early warning steadily to the enterprise that has migration risk through predetermined enterprise migration risk monitoring model, need not to rely on artificial experience, shorten the time of enterprise migration risk before tax control, improve timeliness and the accuracy of losing the early warning response to key tax sources.

Fig. 3 is a schematic structural diagram of a tax fund management apparatus for migrating inauguration enterprises according to an embodiment of the present disclosure, where the apparatus may be understood as the computer device or a part of functional modules in the computer device. As shown in fig. 3, the tax fund management apparatus 300 of the migrating inauguration enterprise may include:

an obtaining module 310, configured to obtain business data of a target enterprise;

the extraction module 320 is used for performing feature extraction processing on the operation data to obtain features related to enterprise migration behaviors of the target enterprise;

and the detection module 330 is configured to input the characteristics into a preset enterprise migration risk monitoring model, and perform migration risk detection on the target enterprise based on the operation data.

Optionally, the tax fund management apparatus 300 for migrating inauguration enterprises may include:

and the preprocessing module is used for preprocessing the operation data, and the preprocessing mode at least comprises one of missing value processing, abnormal value processing, dimension processing, de-duplication processing and noise processing.

Optionally, the characteristics related to the enterprise emigration behavior include at least one of an office supply purchase abnormal index, a supplier regional distribution abnormal index, a house rental abnormal index, a commercial water consumption abnormal index, a commercial electricity consumption abnormal index, a commercial trip cost abnormal index, a recruitment abnormal index, an emigration information index, an enterprise major strategy cooperation event frequency, an industry matching index of a region in which the target enterprise is located, a place of business lease index, and an enterprise legal registration new enterprise index.

Optionally, the operation data includes money data of office supplies purchased by the target enterprise;

the extracting module 320 may include:

the first determining submodule is used for determining a first purchasing amount of the target enterprise for purchasing office supplies every day within preset days based on the amount data of the office supplies purchased by the target enterprise, and calculating an average purchasing amount corresponding to the first purchasing amount within the preset days;

the first calculation submodule is used for calculating the standard deviation of the purchasing amount of office supplies purchased by the target enterprise within the preset days based on the first purchasing amount and the average purchasing amount;

the second calculation submodule is used for calculating the square of a first ratio of a first difference value and an amount standard deviation based on a first difference value of a second purchasing amount of the target enterprise for purchasing office supplies every day and the average purchasing amount within the target days;

the first summing submodule is used for summing the squares of the first specific value corresponding to each day in the target number of days to obtain a first square sum of the target number of days, and the first square sum is determined as an office supply purchase abnormity index of the target enterprise in the target number of days;

and the second determining submodule is used for determining the office supply purchase abnormity index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

Optionally, the business data includes a total invoicing amount of the supplier received by the target enterprise every day, invoicing amounts of suppliers of various jurisdictions received by the target enterprise every day, and supplier data of the target enterprise;

the extracting module 320 may include:

the third calculation operator module is used for calculating a second ratio of the invoicing amount received by the target enterprise in each jurisdiction every day to the total invoicing amount received by the target enterprise every day based on the total invoicing amount of the supplier received by the target enterprise every day and the invoicing amount of the supplier in each jurisdiction every day received by the target enterprise every day;

the fourth calculation submodule is used for calculating the supplier region distribution entropy corresponding to the target enterprise in each day within the preset number of days and the first average distribution entropy corresponding to the supplier region distribution entropy within the preset number of days based on the second ratio;

the fifth calculation submodule is used for calculating a first distribution entropy standard deviation of the supplier region distribution entropy within the preset days based on the supplier region distribution entropy and the first average distribution entropy;

the sixth calculation submodule is used for calculating the absolute value of a second difference value between the provider region distribution entropy and the first average distribution entropy of the target enterprise in each day within the target days, and calculating a third ratio of the absolute value of the second difference value to the standard deviation of the first distribution entropy;

the second summation submodule is used for carrying out summation processing on a third ratio corresponding to each day in the target days to obtain a third ratio sum in the target days, and determining the third ratio sum as a supplier regional distribution abnormal index of the target enterprise in the target days;

and the third determining submodule is used for determining the provider regional distribution abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise emigration behavior.

Optionally, the operation data includes house lease invoice data of the target enterprise;

the extracting module 320 may include:

the fourth determining submodule is used for determining whether the target enterprise has a continuous renting delay behavior of the operating place in the target days or not based on the house renting invoice data of the target enterprise;

a fifth determining submodule, configured to determine a house lease exception index of the target enterprise within the target number of days based on whether the target enterprise has a place renewal postponing behavior within the target number of days;

and the sixth determining submodule is used for determining the house leasing abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise emigration behavior.

Optionally, the business data includes business water amount data of the target enterprise;

the extracting module 320 may include:

the seventh determining submodule is used for determining the first commercial water amount of the target enterprise in each day in preset days based on the commercial water amount data of the target enterprise, and calculating the average water amount corresponding to the first commercial water amount in the preset days;

the seventh calculation submodule is used for calculating the standard deviation of the water consumption amount of the commercial water of the target enterprise in the preset number of days on the basis of the first commercial water consumption amount and the average water consumption amount;

an eighth calculation submodule, configured to calculate, based on a third difference between the second commercial water use amount and the average water use amount of the target enterprise per day in the target number of days, a square of a fourth ratio of the third difference to the standard deviation of the water use amount;

the third summation submodule is used for summing the squares of a fourth ratio corresponding to each day in the target days to obtain a second square sum in the target days, and the second square sum is determined as the business water amount abnormal index of the target enterprise in the target days;

and the eighth determining submodule is used for determining the business water amount abnormal index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise emigration behavior.

Optionally, the operation data includes commercial electricity consumption amount data of the target enterprise;

the extracting module 320 may include:

the ninth determining submodule is used for determining the first commercial electricity utilization amount of the target enterprise in each day within the preset number of days based on the commercial electricity utilization amount data of the target enterprise, and calculating the average electricity utilization amount corresponding to the first commercial electricity utilization amount within the preset number of days;

the ninth calculation submodule is used for calculating the standard deviation of the electricity consumption amount of the commercial electricity of the target enterprise in the preset number of days on the basis of the first commercial electricity consumption amount and the average electricity consumption amount;

a tenth calculation submodule, configured to calculate, based on a fourth difference between the second commercial electricity consumption amount and the average electricity consumption amount of the target enterprise each day within the target number of days, a square of a fifth ratio of the fourth difference to the standard deviation of the electricity consumption amount;

the fourth summation submodule is used for summing the squares of a fifth ratio corresponding to each day in the target number of days to obtain a third square sum in the target number of days, and the third square sum is determined as the commercial power consumption abnormal index of the target enterprise in the target number of days;

and the tenth determining submodule is used for determining the abnormal index of the commercial electricity consumption amount of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise emigration behavior.

Optionally, the business data includes business trip expense data of the target enterprise;

the extracting module 320 may include:

the eleventh determining submodule is used for determining first commercial trip cost of the target enterprise in each day within preset days based on the commercial trip cost data of the target enterprise, and calculating average commercial trip cost corresponding to the first commercial trip cost within the preset days;

the eleventh calculating submodule is used for calculating the standard deviation of the business trip cost of the target enterprise on the business trip within the preset number of days based on the first business trip cost and the average business trip cost;

a twelfth calculating submodule, configured to calculate, based on a fifth difference between a second business trip cost of the target enterprise per day and the average business trip cost within the target number of days, a square of a sixth ratio of the fifth difference to the standard deviation of the business trip costs;

the fifth summation submodule is used for summing the squares of the sixth specific value corresponding to each day in the target days to obtain a fourth square sum in the target days, and the fourth square sum is determined as the business trip cost abnormal index of the target enterprise in the target days;

and the twelfth determining submodule is used for determining the business trip cost abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

Optionally, the operation data includes a total number of target recruitment persons for target recruitment of the target enterprise every day and a number of target recruitment persons for target recruitment of the target enterprise in each jurisdiction every day;

the extracting module 320 may include:

a thirteenth calculating submodule, configured to calculate a seventh ratio of the number of recruiting persons in each jurisdiction per day to the total number of recruiting persons in each day of target recruitment based on the total number of recruiting persons in each jurisdiction per day of target recruitment of the target enterprise and the number of recruiting persons in each jurisdiction per day of target recruitment of the target enterprise within a preset number of days;

the fourteenth calculating submodule is used for calculating the regional distribution entropy of the recruiters of the target enterprises in each day of the target recruitment within the preset days and the second average distribution entropy corresponding to the regional distribution entropy of the recruiters within the preset days based on the seventh ratio;

the fifteenth calculation submodule is used for calculating a second distribution entropy standard deviation of the regional distribution entropy of the recruiters within the preset days based on the regional distribution entropy of the recruiters and the second average distribution entropy;

the sixteenth calculation submodule is used for calculating the absolute value of a sixth difference value between the regional distribution entropy of the recruiters and the second average distribution entropy of the target enterprise in each day within the target days, and calculating an eighth ratio of the sixth difference value to the standard deviation of the second distribution entropy;

the sixth summation submodule is used for carrying out summation processing on an eighth ratio corresponding to each day in the target days to obtain an eighth ratio sum in the target days, and determining the eighth ratio sum as a recruitment abnormal index of the target enterprise in the target days;

and the thirteenth determining submodule is used for determining the recruitment abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise emigration behavior.

Optionally, the operation data includes the enterprise migration information amount of the target enterprise;

the extracting module 320 may include:

the seventh summation submodule is used for summing the target migration information quantity of the target enterprise in each day in the target days based on the enterprise migration information quantity of the target enterprise to obtain the total migration information quantity in the target days;

a fourteenth determining submodule, configured to determine the total migration information quantity as a migration information index of the target enterprise within the target days;

and the fifteenth determining submodule is used for determining the migration information index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

Optionally, the operation data includes event frequency of strategic cooperation between the target enterprise and government organizations in other jurisdictions except the jurisdiction where the target enterprise is located;

the extracting module 320 may include:

the eighth summing sub-module is used for summing the event frequency of each day in the target days to obtain the total event frequency in the target days based on the event frequency of strategic cooperation between the target enterprise and government organizations in other jurisdictions except the jurisdiction where the target enterprise is located;

a sixteenth determining submodule, configured to determine the total event frequency as an enterprise major strategy collaborative event frequency of the target enterprise within the target number of days;

and a seventeenth determining submodule, configured to determine the business major strategy collaborative event frequency of the target enterprise within the target days as a characteristic of the target enterprise related to the enterprise migration behavior.

Optionally, the business data includes upstream and downstream enterprise data of the target enterprise;

the extracting module 320 may include:

a seventeenth calculation submodule for calculating a first fraction of the number of foreign businesses in the upstream business of the target business and a second fraction of the number of foreign businesses in the downstream business of the target business in the number of downstream businesses based on the upstream and downstream business data of the target business;

an eighteenth calculation sub-module for calculating a harmonic mean of the first and second ratios;

an eighteenth determining submodule, configured to determine the harmonic mean as an industry matching index of the jurisdiction where the target enterprise is located in the target days;

and the nineteenth determining submodule is used for determining the industry matching index of the target enterprise in the district in which the target enterprise is located within the target days as the characteristics of the target enterprise related to the enterprise emigration behavior.

Optionally, the operation data includes leasing information of the target enterprise operation place;

the extracting module 320 may include:

a twentieth determining submodule for determining whether the renting information of the target enterprise business place exists in the target days or not based on the renting information of the target enterprise business place;

a twenty-first determining submodule, configured to determine, based on whether rental information of the business place of the target enterprise exists in the target number of days, a business place rental index of the target enterprise in the target number of days;

and a twenty-second determining submodule, configured to determine the place rental index of the target enterprise within the target days as a feature of the target enterprise related to the enterprise migration behavior.

Optionally, the operation data includes registered new enterprise behavior data of enterprise legal persons of the target enterprise in other jurisdictions except the jurisdiction where the target enterprise is located;

the extracting module 320 may include:

a twenty-third determining submodule, configured to determine, based on new enterprise behavior data registered by the enterprise legal person of the target enterprise in other jurisdictions outside the jurisdiction where the target enterprise is located, whether a behavior of registering the new enterprise exists in the enterprise legal person of the target enterprise in other jurisdictions outside the jurisdiction where the target enterprise is located within the target days;

a twenty-fourth determining submodule, configured to determine that the enterprise legal person of the target enterprise registers a new enterprise index within the target number of days based on whether a behavior of registering the new enterprise exists in other jurisdictions outside the jurisdiction where the target enterprise is located by the enterprise legal person of the target enterprise within the target number of days;

and the twenty-fifth determining submodule is used for determining the new enterprise index of the enterprise legal registration of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise emigration behavior.

Optionally, the enterprise emigration risk monitoring model includes a model constructed based on a gradient lifting tree.

Optionally, the enterprise emigration risk monitoring model is a model obtained by training based on an active learning strategy.

The tax fund management device for migrating risky enterprises provided by the embodiment of the disclosure may implement the method of any of the embodiments, and the execution manner and the beneficial effects thereof are similar and will not be described herein again.

The embodiments of the present disclosure further provide a computer device, where the computer device includes a processor and a memory, where the memory stores a computer program, and when the computer program is executed by the processor, the method of any of the embodiments may be implemented, and an execution manner and beneficial effects of the method are similar, which are not described herein again.

The computer device in the embodiments of the present disclosure may be understood as any device having processing and computing capabilities, which may include, but is not limited to, mobile terminals such as smart phones, notebook computers, personal Digital Assistants (PDAs), tablet computers (PADs), portable Multimedia Players (PMPs), vehicle mounted terminals (e.g., car navigation terminals), wearable devices, and the like, as well as stationary electronic devices such as digital TVs, desktop computers, smart home devices, and the like.

Fig. 4 is a schematic structural diagram of a computer device according to an embodiment of the present disclosure, and as shown in fig. 4, the computer device 400 may include a processor 410 and a memory 420, where the memory 420 stores a computer program 421, and when the computer program 421 is executed by the processor 410, the method according to any of the embodiments may be implemented, and the execution manner and the beneficial effects are similar, and are not described again here.

Of course, for simplicity, only some of the components of the computer apparatus 400 relevant to the present invention are shown in fig. 4, and components such as buses, input/output interfaces, input devices, and output devices are omitted. In addition, computer device 400 may include any other suitable components depending on the particular application.

The embodiments of the present disclosure provide a computer-readable storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by a processor, the method of any of the embodiments can be implemented, and the execution manner and the beneficial effects are similar, and are not described herein again.

The computer-readable storage medium described above may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer programs described above may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + +, or the like, as well as conventional procedural programming languages, such as the "C" programming language or similar programming languages, for performing the operations of embodiments of the present disclosure. The program code may execute entirely on the user's computer device, partly on the user's device, as a stand-alone software package, partly on the user's computer device and partly on a remote computer device, or entirely on the remote computer device or server.

The foregoing are merely exemplary embodiments of the present disclosure, which enable those skilled in the art to understand or practice the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (20)

1. A tax fund management method for migrating risky enterprises is characterized by comprising the following steps:

acquiring the operation data of a target enterprise;

performing feature extraction processing on the operation data to obtain features related to the migration behavior of the target enterprise and the enterprise;

inputting the characteristics into a preset enterprise migration risk monitoring model, and detecting the migration risk of the target enterprise based on the enterprise migration risk monitoring model.

2. The method of claim 1, wherein prior to performing the feature extraction process on the business data, the method further comprises:

and preprocessing the operation data, wherein the preprocessing mode at least comprises one of missing value processing, abnormal value processing, dimension processing, duplication removing processing and noise processing.

3. The method of claim 1, wherein the characteristics related to the enterprise migration behavior comprise at least one of an office supply procurement anomaly index, a supplier regional distribution anomaly index, a house rental anomaly index, a commercial water use amount anomaly index, a commercial electricity use amount anomaly index, a commercial trip cost anomaly index, a recruitment anomaly index, a migration information index, an enterprise major strategy cooperation event frequency, an industry matching index of a jurisdiction where the target enterprise is located, a place of business lease index, and an enterprise legal registration new enterprise index.

4. The method of claim 3, wherein the business data includes data on the amount of office supplies purchased by the target enterprise;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining a first purchase amount of the target enterprise for purchasing office supplies every day within a preset number of days based on the amount data of the office supplies purchased by the target enterprise, and calculating an average purchase amount corresponding to the first purchase amount within the preset number of days;

calculating the standard deviation of the purchasing amount of office supplies purchased by the target enterprise within the preset days based on the first purchasing amount and the average purchasing amount;

calculating a square of a first ratio of a second difference value to the standard deviation of the procurement amount based on a first difference value between a second procurement amount for the target enterprise to procure office supplies per day and the average procurement amount within a target number of days;

summing the squares of the first ratio corresponding to each day in the target number of days to obtain a first square sum of the target number of days, and determining the first square sum as an office supply purchase abnormity index of the target enterprise in the target number of days;

and determining the office supply procurement abnormity index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

5. The method of claim 3, wherein said business data includes a total invoiced amount of suppliers received by said target enterprise on a daily basis and an invoiced amount of suppliers for each jurisdiction received by said target enterprise on a daily basis, and supplier data for said target enterprise;

the step of performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise comprises the following steps:

calculating a second ratio of the invoicing amount received by the target enterprise in each jurisdiction every day to the total invoicing amount received by the target enterprise every day based on the total invoicing amount of the supplier received by the target enterprise every day and the invoicing amount of the supplier in each jurisdiction received by the target enterprise every day;

calculating a supplier region distribution entropy corresponding to the target enterprise every day within the preset number of days and a first average distribution entropy corresponding to the supplier region distribution entropy within the preset number of days based on the second ratio;

calculating a first distribution entropy standard deviation of the supplier region distribution entropy within the preset number of days based on the supplier region distribution entropy and the first average distribution entropy;

calculating an absolute value of a second difference value between the supplier region distribution entropy and the first average distribution entropy of the target enterprise every day within a target number of days, and calculating a third ratio of the absolute value of the second difference value to the standard deviation of the first distribution entropy;

summing the third ratio corresponding to each day in the target days to obtain a third ratio sum in the target days, and determining the third ratio sum as a supplier regional distribution abnormal index of the target enterprise in the target days;

and determining the provider regional distribution abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

6. The method of claim 3, wherein the business data comprises house rental invoice data for the target business;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining whether the target enterprise has a management place renewal postponing action within the target number of days or not based on the house lease invoice data of the target enterprise;

determining a house lease abnormal index of the target enterprise in the target days based on whether the target enterprise has the continuous lease postponing behavior of the business place in the target days;

and determining the house leasing abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

7. The method of claim 3, wherein the business data includes business water usage amount data for the target enterprise;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining a first commercial water amount of the target enterprise in a preset number of days based on the commercial water amount data of the target enterprise, and calculating an average water amount corresponding to the first commercial water amount in the preset number of days;

calculating the standard deviation of the water consumption amount of the commercial water of the target enterprise in the preset days on the basis of the first commercial water consumption amount and the average water consumption amount;

calculating a square of a fourth ratio of a third difference to the standard deviation of the water usage amount based on a third difference between a second business water usage amount and the average water usage amount for each day of the target business over the target number of days;

summing the squares of the fourth ratio corresponding to each day in the target number of days to obtain a second square sum in the target number of days, and determining the second square sum as the business water amount abnormal index of the target enterprise in the target number of days;

and determining the business water amount abnormal index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise emigration behavior.

8. The method of claim 3, wherein the business data includes business electricity amount data for the target enterprise;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining a first commercial electricity consumption amount of the target enterprise every day in a preset number of days based on the commercial electricity consumption amount data of the target enterprise, and calculating an average electricity consumption amount corresponding to the first commercial electricity consumption amount in the preset number of days;

calculating the standard deviation of the electricity consumption amount of the commercial electricity of the target enterprise in the preset days based on the first commercial electricity consumption amount and the average electricity consumption amount;

calculating a square of a fifth ratio of a fourth difference to the standard deviation of the electricity usage amount based on a fourth difference between a second commercial electricity usage amount for the target business each day over the target number of days and the average electricity usage amount;

summing the squares of the fifth specific value corresponding to each day in the target number of days to obtain a third square sum in the target number of days, and determining the third square sum as the business electricity sum abnormal index of the target enterprise in the target number of days;

and determining the abnormal index of the commercial electricity consumption amount of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

9. The method of claim 3, wherein the business data comprises business trip cost data for the target enterprise;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining first business trip cost of the target enterprise in preset days each day based on the business trip cost data of the target enterprise, and calculating average business trip cost corresponding to the first business trip cost in the preset days;

calculating the standard deviation of the business trip cost of the target enterprise on the business trip within the preset number of days based on the first business trip cost and the average business trip cost;

calculating a square of a sixth ratio of a fifth difference to the standard deviation of business trip costs based on a fifth difference between a second business trip cost of the target business per day and the average business trip cost over a target number of days;

summing the squares of the sixth ratio corresponding to each day in the target number of days to obtain a fourth square sum in the target number of days, and determining the fourth square sum as the business trip cost abnormal index of the target enterprise in the target number of days;

and determining the business travel cost abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

10. The method of claim 3, wherein the business data includes a total number of recruits for each day of targeted recruitment for the target enterprise and a number of recruiters for each day of targeted recruitment for the target enterprise in each jurisdiction;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

calculating a seventh ratio of the number of recruiting persons for each target recruitment of the target enterprise in each jurisdiction to the total number of recruiting persons for each target recruitment of the target enterprise in each jurisdiction within a preset number of days based on the total number of recruiting persons for each target recruitment of the target enterprise and the number of recruiting persons for each target recruitment of the target enterprise in each jurisdiction every day;

calculating a regional distribution entropy of the recruiters for target recruitment every day of the target enterprise within the preset days and a second average distribution entropy corresponding to the regional distribution entropy of the recruiters within the preset days based on the seventh ratio;

calculating a second distribution entropy standard deviation of the regional distribution entropy of the recruiters in the preset days based on the regional distribution entropy of the recruiters and the second average distribution entropy;

calculating an absolute value of a sixth difference value between the regional distribution entropy of the recruiters and the second average distribution entropy of the target enterprise in each day within the target days, and calculating an eighth ratio of the sixth difference value to the standard deviation of the second distribution entropy;

summing the eighth ratio corresponding to each day in the target days to obtain an eighth ratio sum in the target days, and determining the eighth ratio sum as a recruitment abnormity index of the target enterprise in the target days;

and determining the recruitment abnormal index of the target enterprise in the target days as the characteristic of the target enterprise related to the enterprise migration behavior.

11. The method of claim 3, wherein the business data includes a business migratory information volume for the target business;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

summing the target migration information quantity of the target enterprise in each day in the target number of days to obtain the total migration information quantity in the target number of days;

determining the total migration information quantity as the migration information index of the target enterprise in the target days;

and determining the migration information index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

12. The method of claim 3, wherein the business data comprises event frequency for strategic collaboration between the target enterprise and governmental organizations outside the jurisdiction in which the target enterprise is located;

the step of performing feature extraction processing on the operation data to obtain features related to enterprise emigration behaviors of the target enterprise comprises the following steps:

based on the event frequency of strategic cooperation between the target enterprise and other government organizations outside the jurisdiction where the target enterprise is located, summing the event frequency of each day in the target number of days to obtain the total event frequency in the target number of days;

determining the total event frequency as the business big strategy collaborative event frequency of the target business in the target days;

and determining the business major strategy cooperative event frequency of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

13. The method of claim 3, wherein the business data comprises business data upstream and downstream of the target business;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

calculating a first proportion of the number of foreign businesses in an upstream business of the target business to the number of upstream businesses and a second proportion of the number of foreign businesses in a downstream business of the target business to the number of downstream businesses within a target number of days based on upstream and downstream business data of the target business;

calculating a harmonic mean of the first and second ratios;

determining the harmonic mean as an industry matching index of the target enterprise in the district in which the target enterprise is located within the target days;

and determining the industry matching index of the target enterprise in the district in which the target enterprise is located in the target days as the characteristic related to the enterprise migration behavior of the target enterprise.

14. The method of claim 3, wherein the business data includes rental information for the target enterprise business;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining whether the renting information of the target enterprise business place exists in target days or not based on the renting information of the target enterprise business place;

determining the operating place renting index of the target enterprise in the target number of days based on whether the renting information of the target enterprise operating place exists in the target number of days;

and determining the business place lease index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise migration behavior.

15. The method of claim 3, wherein said operational data comprises registered new enterprise behavior data of enterprise jurisdictions of said target enterprise outside of said jurisdiction where said target enterprise is located;

the characteristic extraction processing is carried out on the operation data to obtain the characteristics related to the migration behavior of the target enterprise and the enterprise, and the method comprises the following steps:

determining whether the enterprise legal person of the target enterprise has the behavior of registering the new enterprise in other jurisdictions outside the jurisdiction of the target enterprise within the target days based on the registered new enterprise behavior data of the enterprise legal person of the target enterprise in other jurisdictions outside the jurisdiction of the target enterprise;

determining an enterprise legal person registration new enterprise index of the target enterprise in the target day number based on whether the enterprise legal person of the target enterprise has a new enterprise registration behavior in other jurisdictions except the jurisdiction where the target enterprise is located in the target day number;

and determining the enterprise legal registration new enterprise index of the target enterprise in the target days as the characteristics of the target enterprise related to the enterprise emigration behavior.

16. The method of claim 1, wherein the enterprise migration risk monitoring model comprises a model constructed based on a gradient lift tree.

17. The method of claim 1, wherein the enterprise emigration risk monitoring model is a model trained based on an active learning strategy.

18. The utility model provides a move out tax fund management device of inauguration enterprise which characterized in that includes:

the acquisition module is used for acquiring the operation data of the target enterprise;

the extraction module is used for carrying out feature extraction processing on the operation data to obtain features related to the migration behavior of the target enterprise and the enterprise;

and the detection module is used for inputting the characteristics into a preset enterprise migration risk monitoring model and detecting the migration risk of the target enterprise based on the enterprise migration risk monitoring model.

19. A computer device, comprising:

memory and a processor, wherein the memory has stored therein a computer program which, when executed by the processor, implements the tax fund management method of an explanted inauguration enterprise according to any of the claims 1-17.

20. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, implements the tax fund management method for migrating inauguration enterprises according to any one of claims 1 to 17.

Images