CN110264342A - A kind of business audit method and device based on machine learning - Google Patents
A kind of business audit method and device based on machine learning Download PDFInfo
- Publication number
- CN110264342A CN110264342A CN201910533825.2A CN201910533825A CN110264342A CN 110264342 A CN110264342 A CN 110264342A CN 201910533825 A CN201910533825 A CN 201910533825A CN 110264342 A CN110264342 A CN 110264342A
- Authority
- CN
- China
- Prior art keywords
- random forest
- forest model
- decision tree
- audit
- classification results
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012550 audit Methods 0.000 title claims abstract description 111
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000010801 machine learning Methods 0.000 title claims abstract description 23
- 238000007637 random forest analysis Methods 0.000 claims abstract description 89
- 238000003066 decision tree Methods 0.000 claims abstract description 85
- 238000012549 training Methods 0.000 claims abstract description 35
- 238000012545 processing Methods 0.000 claims description 18
- 238000003860 storage Methods 0.000 claims description 10
- 235000013399 edible fruits Nutrition 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 7
- 230000006399 behavior Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 14
- 238000004590 computer program Methods 0.000 description 10
- 238000005034 decoration Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 3
- 238000011144 upstream manufacturing Methods 0.000 description 3
- 230000003068 static effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The business audit method and device based on machine learning that the embodiment of the invention provides a kind of, it is related to financial technology technical field, this method comprises: using history service audit data training Random Forest model, make the audit behavior of Random Forest model study auditor, then by every decision tree in the feature vector input Random Forest model of business applicant, obtain the classification results of every decision tree output in Random Forest model, the corresponding weight of every decision tree in the classification results and Random Forest model exported later further according to every decision tree in Random Forest model, determine the auditing result of service request.For expert model, Random Forest model is obtained based on history service audit data training, is not merely the experience of professional auditor, therefore rely on artificial experience small, the influence for reducing subjective factor improves the generalization ability and versatility of audit model.
Description
Technical field
The present embodiments relate to financial technology (Fintech) technical fields more particularly to a kind of based on machine learning
Business audit method and device.
Background technique
With the development of computer technology, and more and more technical applications (such as: artificial intelligence, cloud computing, block chain
Deng) in financial field, traditional financial industry is gradually changing to financial technology (Fintech), but due to the safety of financial industry
Property, requirement of real-time, also to technology propose higher requirement.In financial industry, supply chain loan solves upstream and downstream enterprise
The difficult problem of industry financing difficulties, guarantee, and by getting through upstream and downstream financing bottleneck, supply chain finance costs can also be reduced,
Improve the competitiveness of core enterprise and mating enterprise.Currently, mainly loan is audited using expert model, it is more by collecting
It is engaged in the experience of the business personnel of loan audit year, by summarizing, is summarized as the service logic rule of a set of fixation, passes through
Regulation engine is deployed in auditing system, realizes automatic audit, and this method excessively relies on the experience of business personnel, subjective factor
By force.
Summary of the invention
Due to excessively relying on the experience of business personnel using the scheme of expert model audit business, subjective factor is strong to be asked
Topic, the business audit method and device based on machine learning that the embodiment of the invention provides a kind of.
On the one hand, the business audit method based on machine learning that the embodiment of the invention provides a kind of, comprising:
The service request of acquisition business applicant；
The feature vector of the business applicant is extracted according to the service request；
By every decision tree in the feature vector input Random Forest model of the business applicant, obtain described random
The classification results that every decision tree exports in forest model, the Random Forest model are with history service audit data for training
What sample training obtained；
According to the classification results that every decision tree in the Random Forest model exports, the audit of the service request is determined
As a result.
Optionally, the classification results exported according to every decision tree in the Random Forest model, determine the industry
The auditing result of business request, comprising:
It is every in the classification results and the Random Forest model exported according to every decision tree in the Random Forest model
The corresponding weight of decision tree, determines the auditing result of the service request.
Optionally, described according to the classification results of every decision tree output in the Random Forest model and described random gloomy
The corresponding weight of every decision tree in woods model, determines the auditing result of the service request, comprising:
The weight of the identical decision tree of classification results in the Random Forest model is added, determines each classification results
Classified weight；
Using the maximum classification results of classified weight as auditing result.
Optionally, the service request is supply chain service request, and the history service audit data include Lian Shu enterprise
Characteristic, core enterprise's characteristic, auditor history audit logging.
Optionally, the Random Forest model is obtained by training sample training of history service audit data, comprising:
It obtains history service and audits data；
Data, which are audited, according to the history service determines feature vector set；
N number of subcharacter vector set is extracted from described eigenvector set, the N is default positive integer；
N decision tree is obtained using N number of subcharacter vector set training；
N decision tree is formed into Random Forest model.
On the one hand, the embodiment of the invention provides a kind of business audit device based on machine learning, comprising:
Module is obtained, for obtaining the service request of business applicant；
Extraction module, for extracting the feature vector of the business applicant according to the service request；
Categorization module, for every decision in the feature vector input Random Forest model by the business applicant
Tree obtains the classification results that every decision tree exports in the Random Forest model, and the Random Forest model is with history industry
Business audit data are that training sample training obtains；
Processing module, for according in the Random Forest model every decision tree export classification results, determine described in
The auditing result of service request.
Optionally, the processing module is specifically used for:
It is every in the classification results and the Random Forest model exported according to every decision tree in the Random Forest model
The corresponding weight of decision tree, determines the auditing result of the service request.
Optionally, the processing module is specifically used for:
The weight of the identical decision tree of classification results in the Random Forest model is added, determines each classification results
Classified weight；
Using the maximum classification results of classified weight as auditing result.
Optionally, the service request is supply chain service request, and the history service audit data include Lian Shu enterprise
Characteristic, core enterprise's characteristic, auditor history audit logging.
Optionally, the categorization module is specifically used for:
It obtains history service and audits data；
Data, which are audited, according to the history service determines feature vector set；
N number of subcharacter vector set is extracted from described eigenvector set, the N is default positive integer；
N decision tree is obtained using N number of subcharacter vector set training；
N decision tree is formed into Random Forest model.
On the one hand, the embodiment of the invention provides a kind of computer equipment, including memory, processor and it is stored in storage
On device and the computer program that can run on a processor, the processor are realized when executing described program based on machine learning
The step of business audit method.
On the one hand, the embodiment of the invention provides a kind of computer readable storage medium, being stored with can be set by computer
The standby computer program executed, when described program is run on a computing device, so that computer equipment execution is based on
The step of business audit method of machine learning.
In the embodiment of the present invention, due to making random forest mould using history service audit data training Random Forest model
Type learns the audit behavior of auditor, then by every decision in the feature vector input Random Forest model of business applicant
Tree obtains the classification results of every decision tree output in Random Forest model, later certainly further according to every in Random Forest model
The classification results of plan tree output, determine the auditing result of service request.For expert model, Random Forest model is base
It is obtained in history service audit data training, is not merely the experience of professional auditor, therefore small to the dependence of artificial experience,
The influence for reducing subjective factor improves the generalization ability and versatility of audit model.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill in field, without any creative labor, it can also be obtained according to these attached drawings
His attached drawing.
Fig. 1 is a kind of application scenarios schematic diagram provided in an embodiment of the present invention；
Fig. 2 is a kind of flow diagram of the business audit method based on machine learning provided in an embodiment of the present invention；
Fig. 3 is a kind of flow diagram of the method for trained Random Forest model provided in an embodiment of the present invention；
Fig. 4 is a kind of flow diagram of the business audit method based on machine learning provided in an embodiment of the present invention；
Fig. 5 is a kind of structural schematic diagram of the business audit device based on machine learning provided in an embodiment of the present invention；
Fig. 6 is a kind of structural schematic diagram of computer equipment provided in an embodiment of the present invention.
Specific embodiment
In order to which the purpose of the present invention, technical solution and beneficial effect is more clearly understood, below in conjunction with attached drawing and implementation
Example, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is only used to explain this hair
It is bright, it is not intended to limit the present invention.
In order to facilitate understanding, noun involved in the embodiment of the present invention is explained below.
Supply chain loan: Supply Chain Finance, be in supply chain core enterprise and its it is relevant up and down
You Lianshu enterprise as a whole, formulates according to the transaction relationship of enterprise in supply chain and industry characteristic based on goods power and cash
A kind of Financing Mode of the whole finance resolution of flow control.
Core enterprise: in supply chain grasp core technology, core competence, core link enterprise.
Lian Shu enterprise: the upstream and downstream Lian Shu enterprise of core enterprise in supply chain.
The business audit method based on machine learning in the embodiment of the present invention can be applied to applied field as shown in Figure 1
Scape includes terminal device 101, audit server 102 in the application scenarios, wherein terminal device 101 can be intelligent hand
Machine, tablet computer or portable personal computer etc..Audit server 102 can be the business audit of bank and other financial mechanism
Server.User submits service request on terminal device 101, and service request is sent to audit server by terminal device 101
102.Auditing includes the trained Random Forest model for business audit in server 102.Server 102 is audited according to institute
Service request extracts the feature vector of business applicant, will be every in the feature vector input Random Forest model of business applicant
Decision tree obtains the classification results of every decision tree output in Random Forest model, certainly according to every in Random Forest model
The classification results of plan tree output, determine the auditing result of service request.Auditing result is sent to terminal and set by audit server 102
Standby 101, user can check the auditing result of service request from terminal device 101.
Based on application scenario diagram shown in FIG. 1, the embodiment of the invention provides a kind of business audits based on machine learning
The process of the process of method, this method can be executed by the business audit device based on machine learning, the industry based on machine learning
Business audit device can be the audit server 102 in Fig. 1, as shown in Figure 2, comprising the following steps:
Step S201 obtains the service request of business applicant.
Step S202 extracts the feature vector of business applicant according to service request.
Business applicant can be personal or enterprise, and service request can be loan requests, such as personal loan request, enterprise
Industry loan requests, supply chain loan requests etc..
When business applicant is personal, the feature vector of business applicant can be personal characteristics, such as personal basic
Information, personal reference record, personal business record, personal asset etc..When business applicant is enterprise, the spy of business applicant
Sign vector can be enterprise characteristic, such as enterprise's reference record, enterprise qualification etc..
Step S203, by every decision tree in the feature vector of business applicant input Random Forest model, obtain with
The classification results that every decision tree exports in machine forest model.
Random Forest model is obtained by training sample training of history service audit data.When Random Forest model is used
When auditing personal service request, history service audit data include the history audit logging of personal characteristics data, auditor.When
Random Forest model is for when auditing supply chain service request, it to include chain category enterprise characteristic data, core that history service, which audits data,
Heart enterprise characteristic data, auditor history audit logging.
Step S204 determines examining for service request according to the classification results that every decision tree in Random Forest model exports
Core result.
In a kind of possible embodiment, the classification results of decision tree output include auditing to pass through and audit not passing through,
When the classification results exported in Random Forest model are that the decision tree quantity that audit passes through is greater than the unacceptable decision tree number of audit
When amount, determine that the auditing result of service request passes through for audit.When the classification results exported in Random Forest model are that audit is logical
When the decision tree quantity crossed is less than audit unacceptable decision tree quantity, it is obstructed to audit to determine the auditing result of service request
It crosses.
In a kind of possible embodiment, according to the classification results of every decision tree output in Random Forest model and with
The corresponding weight of every decision tree in machine forest model, determines the auditing result of service request.Specifically, each decision tree is corresponding
Weight be to be determined according to the importance of feature vector in decision tree, it is that the corresponding weight of all decision trees is added and be 1.
In specific implementation, the weight of the identical decision tree of classification results in Random Forest model can be added, be determined every
The classified weight of a classification results, using the maximum classification results of classified weight as auditing result.
Illustratively, the classification results for setting decision tree include auditing to pass through and audit not passing through, are careful by classification results
The weight for the decision tree that core passes through is added, and obtains the classified weight that audit passes through.It is to audit unacceptable decision by classification results
The weight of tree is added, and is obtained and is audited unacceptable classified weight.When the classified weight that audit passes through is greater than unacceptable point of audit
It when class weight, determines that auditing result passes through for audit, audits unacceptable classified weight when the classified weight that audit passes through is less than
When, determine that auditing result does not pass through for audit.
Due to making examining for Random Forest model study auditor using history service audit data training Random Forest model
Core behavior obtains random forest then by every decision tree in the feature vector input Random Forest model of business applicant
The classification results that every decision tree exports in model, the classification knot exported later further according to every decision tree in Random Forest model
Fruit determines the auditing result of service request.For expert model, Random Forest model is to audit number based on history service
It is obtained according to training, is not merely the experience of professional auditor, therefore is small to the dependence of artificial experience, reduce subjective factor
It influences, improves the generalization ability and versatility of audit model.
The process that Random Forest model is obtained using history service audit data training is described below, as shown in figure 3, including
Following steps:
Step S301 obtains history service and audits data.
History service audit data include chain category enterprise characteristic data, core enterprise's characteristic, the history of auditor examine
Core record, wherein chain category enterprise characteristic data include Lian Shu enterprise essential information (for example, establish with limited laibility the time, scope of the enterprise,
Headcount, main business industry etc.), chain category enterprise operation financial report (for example, debt ratio, amount of liabilities, the external guaranty amount of money,
The turnover, operating profit etc.), pledged the accounts receivable amount of money, pledged accounts receivable stroke count, accounts receivable characteristic (ratio
Such as accounts receivable account phase).Core enterprise's characteristic includes core enterprise's essential information (for example, establishing time, enterprise with limited laibility
Whether scale headcount, main business industry, is listed company etc.), that core enterprise manages financial report, recent public sentiment is great negative
Face news quantity, accounts receivable characteristic (such as accounts receivable account phase etc.).
Step S302 audits data according to history service and determines feature vector set.
Specifically, after obtaining history service audit data, history service audit data are pre-processed.Pretreatment can be with
Including following methods: mode one, since the history service of collection audits data, there may be mistake, exceptional value and missings
Situation can do the operation such as tax default value, Rejection of samples to this kind of data, avoid this kind of data influence training result.Mode two,
Standardization is done to specific characteristic column, such as the operating income of company is one-dimensional continuously distributed numerical data, can be done herein
Section classification, such as 3,000,000 or less, 300 ten thousand to 1,000 ten thousand, 1,000 ten thousand to 5,000 ten thousand, 50,000,000 or more 4 section classifications, then
It is marked.History service is audited data preparation for multidimensional data matrix, to use to training by mode three.
Step S303, extracts N number of subcharacter vector set from feature vector set, and N is default positive integer.
Specifically, N number of subcharacter vector set is extracted from feature vector set using bootstrap method.
Step S304 obtains N decision tree using the training of N number of subcharacter vector set, and N decision tree composition is random
Forest model.
For each subcharacter vector set, decision is obtained using the feature vector training in the subcharacter vector set
Tree, wherein decision tree can be CART (classification returns) tree.According to the importance of the feature vector in subcharacter vector set,
The weight of decision tree is set, and the sum of weight of N decision tree is 1.
Using history service audit data training Random Forest model, while being arranged at random according to the importance of feature vector
The weight of every decision tree in forest model, therefore when using Random Forest model audit service request, in conjunction with point of decision tree
Class result and the weight of decision tree can effectively provide the accuracy of auditing result.
Embodiment in order to preferably explain the present invention is provided a loan using supply chain below and describes present invention implementation as implement scene
A kind of business audit method based on machine learning that example provides, this method are held by the business audit device based on machine learning
Row, as shown in figure 4, method includes the following steps:
Setting bank A has more loan transaction in supply chain financial field, and the data warehouse of operation system has accumulated confession
Answer the loan enterprises application record and loan audit historical data of chain business.From business system data, row in risk data and
Feature vector set Dt is extracted in people's row collage-credit data.N number of son is extracted from feature vector set Dt using bootstrap method
Feature vector set { D1, D2 ..., DN }.N decision tree is obtained using the training of N number of subcharacter vector set, respectively T1,
T2 ..., TN, according to the importance of the feature vector in subcharacter vector set, the weight that N decision tree is arranged is respectively
{ a1, a2 ..., a N }, the sum of weight of N decision tree are that 1, N decision tree forms Random Forest model.
A property finishing has occurred between C real estate company, core enterprise in supply chain and Lian Shu enterprise B decoration corporation
Business, needs C real estate company after standby service to pay B decoration corporation by 5,000,000 yuan of the business amount of money, and B decoration corporation needs at this time
It wants business funds to have enough to meet the need, passes through the accounts receivable of this business trade background, Xiang Yinhang A application loan three and one-half million member.Bank A
After the loan requests for receiving B decoration corporation, Cong Renhang reference platform inquires the reference record of B decoration corporation, including history is borrowed
Money registration record, pledge record, assets pledge record, enterprise's external guaranty record etc., then from enterprise register in Administration for Industry and Commerce data
Source, obtain C real estate company and B decoration corporation enterprise characteristic data, such as scope of the enterprise, set up the time, registered enterprise fund,
Manage financial report, accounts receivable characteristic etc..The feature of C real estate company is extracted from the data being related to above for this loan
Feature vector is separately input in N decision tree by vector, and every decision tree exports a classification results, wherein classification results
Pass through and audit including audit and does not pass through.The weight that classification results are the decision tree that audit passes through is added, audit is obtained and passes through
Classified weight.It is the weight addition for auditing unacceptable decision tree by classification results, obtains and audit unacceptable classified weight.
When the classified weight that audit passes through, which is greater than, audits unacceptable classified weight, determines that auditing result passes through for audit, work as audit
By classified weight be less than audit unacceptable classified weight when, determine auditing result for audit do not pass through.
Due to making examining for Random Forest model study auditor using history service audit data training Random Forest model
Core behavior obtains random forest then by every decision tree in the feature vector input Random Forest model of business applicant
The classification results that every decision tree exports in model, the classification knot exported later further according to every decision tree in Random Forest model
Fruit determines the auditing result of service request.For expert model, Random Forest model is to audit number based on history service
It is obtained according to training, is not merely the experience of professional auditor, therefore is small to the dependence of artificial experience, reduce subjective factor
It influences, improves the generalization ability and versatility of audit model.
Based on the same technical idea, the embodiment of the invention provides a kind of business audit device based on machine learning,
As shown in figure 5, the device 500 includes:
Module 501 is obtained, for obtaining the service request of business applicant；
Extraction module 502, for extracting the feature vector of the business applicant according to the service request；
Categorization module 503, certainly for every in the feature vector input Random Forest model by the business applicant
Plan tree obtains the classification results that every decision tree exports in the Random Forest model, and the Random Forest model is with history
Business audit data are that training sample training obtains；
Processing module 504, the classification results for being exported according to every decision tree in the Random Forest model, determines institute
State the auditing result of service request.
Optionally, the processing module 504 is specifically used for:
It is every in the classification results and the Random Forest model exported according to every decision tree in the Random Forest model
The corresponding weight of decision tree, determines the auditing result of the service request.
Optionally, the processing module 504 is specifically used for:
The weight of the identical decision tree of classification results in the Random Forest model is added, determines each classification results
Classified weight；
Using the maximum classification results of classified weight as auditing result.
Optionally, the service request is supply chain service request, and the history service audit data include Lian Shu enterprise
Characteristic, core enterprise's characteristic, auditor history audit logging.
Optionally, the categorization module 503 is specifically used for:
It obtains history service and audits data；
Data, which are audited, according to the history service determines feature vector set；
N number of subcharacter vector set is extracted from described eigenvector set, the N is default positive integer；
N decision tree is obtained using N number of subcharacter vector set training；
N decision tree is formed into Random Forest model.
Based on the same technical idea, the embodiment of the invention provides a kind of computer equipments, as shown in fig. 6, including extremely
Lack a processor 601, and the memory 602 connecting at least one processor, does not limit processing in the embodiment of the present invention
Specific connection medium between device 601 and memory 602 passes through bus between processor 601 and memory 602 in Fig. 6 and connects
For.Bus can be divided into address bus, data/address bus, control bus etc..
In embodiments of the present invention, memory 602 is stored with the instruction that can be executed by least one processor 601, at least
The instruction that one processor 601 is stored by executing memory 602, can execute the business audit above-mentioned based on machine learning
Included step in method.
Wherein, processor 601 is the control centre of computer equipment, can use various interfaces and connection computer
The various pieces of equipment are stored in memory 602 by running or executing the instruction being stored in memory 602 and calling
Data, to carry out business audit.Optionally, processor 601 may include one or more processing units, and processor 601 can
Integrated application processor and modem processor, wherein the main processing operation system of application processor, user interface and application
Program etc., modem processor mainly handle wireless communication.It is understood that above-mentioned modem processor can not also
It is integrated into processor 601.In some embodiments, processor 601 and memory 602 can be realized on the same chip,
In some embodiments, they can also be realized respectively on independent chip.
Processor 601 can be general processor, such as central processing unit (CPU), digital signal processor, dedicated integrated
Circuit (Application Specific Integrated Circuit, ASIC), field programmable gate array or other can
Perhaps transistor logic, discrete hardware components may be implemented or execute present invention implementation for programmed logic device, discrete gate
Each method, step and logic diagram disclosed in example.General processor can be microprocessor or any conventional processor
Deng.The step of method in conjunction with disclosed in the embodiment of the present invention, can be embodied directly in hardware processor and execute completion, Huo Zheyong
Hardware and software module combination in processor execute completion.
Memory 602 is used as a kind of non-volatile computer readable storage medium storing program for executing, can be used for storing non-volatile software journey
Sequence, non-volatile computer executable program and module.Memory 602 may include the storage medium of at least one type,
It such as may include flash memory, hard disk, multimedia card, card-type memory, random access storage device (Random Access
Memory, RAM), static random-access memory (Static Random Access Memory, SRAM), may be programmed read-only deposit
Reservoir (Programmable Read Only Memory, PROM), read-only memory (Read Only Memory, ROM), band
Electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory,
EEPROM), magnetic storage, disk, CD etc..Memory 602 can be used for carrying or storing have instruction or data
The desired program code of structure type and can by any other medium of computer access, but not limited to this.The present invention is real
Applying the memory 602 in example can also be circuit or other devices that arbitrarily can be realized store function, for storing program
Instruction and/or data.
Based on the same technical idea, it the embodiment of the invention provides a kind of computer readable storage medium, is stored with
The computer program that can be executed by computer equipment, when described program is run on a computing device, so that the computer
Equipment executes the step of business audit method based on machine learning.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the present invention
Form.It is deposited moreover, the present invention can be used to can be used in the computer that one or more wherein includes computer usable program code
The shape for the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of business audit method based on machine learning characterized by comprising
The service request of acquisition business applicant；
The feature vector of the business applicant is extracted according to the service request；
By every decision tree in the feature vector input Random Forest model of the business applicant, the random forest is obtained
The classification results that every decision tree exports in model, the Random Forest model are using history service audit data as training sample
What training obtained；
According to the classification results that every decision tree in the Random Forest model exports, the audit knot of the service request is determined
Fruit.
2. the method as described in claim 1, which is characterized in that described defeated according to every decision tree in the Random Forest model
Classification results out determine the auditing result of the service request, comprising:
It determines for every in the classification results and the Random Forest model exported according to every decision tree in the Random Forest model
The corresponding weight of plan tree, determines the auditing result of the service request.
3. method according to claim 2, which is characterized in that described defeated according to every decision tree in the Random Forest model
The corresponding weight of every decision tree in classification results and the Random Forest model out, determines the audit knot of the service request
Fruit, comprising:
The weight of the identical decision tree of classification results in the Random Forest model is added, determines the classification of each classification results
Weight；
Using the maximum classification results of classified weight as auditing result.
4. the method as described in claims 1 to 3 is any, which is characterized in that the service request is supply chain service request, institute
State history service audit data include chain category enterprise characteristic data, core enterprise's characteristic, auditor history audit logging.
5. method as claimed in claim 4, which is characterized in that the Random Forest model is to be with history service audit data
Training sample training obtains, comprising:
It obtains history service and audits data；
Data, which are audited, according to the history service determines feature vector set；
N number of subcharacter vector set is extracted from described eigenvector set, the N is default positive integer；
N decision tree is obtained using N number of subcharacter vector set training；
N decision tree is formed into Random Forest model.
6. a kind of business audit device based on machine learning characterized by comprising
Module is obtained, for obtaining the service request of business applicant；
Extraction module, for extracting the feature vector of the business applicant according to the service request；
Categorization module is obtained for every decision tree in the feature vector input Random Forest model by the business applicant
The classification results of every decision tree output in the Random Forest model are obtained, the Random Forest model is audited with history service
Data are that training sample training obtains；
Processing module, the classification results for being exported according to every decision tree in the Random Forest model, determines the business
The auditing result of request.
7. device as claimed in claim 6, which is characterized in that the processing module is specifically used for:
It determines for every in the classification results and the Random Forest model exported according to every decision tree in the Random Forest model
The corresponding weight of plan tree, determines the auditing result of the service request.
8. device as claimed in claim 7, which is characterized in that the processing module is specifically used for:
The weight of the identical decision tree of classification results in the Random Forest model is added, determines the classification of each classification results
Weight；
Using the maximum classification results of classified weight as auditing result.
9. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor is realized described in Claims 1 to 5 any claim when executing described program
The step of method.
10. a kind of computer readable storage medium, which is characterized in that it is stored with the computer journey that can be executed by computer equipment
Sequence, when described program is run on a computing device, so that computer equipment perform claim requirement 1~5 is any described
The step of method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910533825.2A CN110264342B (en) | 2019-06-19 | 2019-06-19 | Business auditing method and device based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910533825.2A CN110264342B (en) | 2019-06-19 | 2019-06-19 | Business auditing method and device based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110264342A true CN110264342A (en) | 2019-09-20 |
| CN110264342B CN110264342B (en) | 2024-06-28 |
Family
ID=67919576
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910533825.2A Active CN110264342B (en) | 2019-06-19 | 2019-06-19 | Business auditing method and device based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110264342B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111612166A (en) * | 2020-04-23 | 2020-09-01 | 中国科学院计算机网络信息中心 | Reimbursement time prediction method based on machine learning |
| CN112258135A (en) * | 2020-05-15 | 2021-01-22 | 北京沃东天骏信息技术有限公司 | Method and device for auditing prescription data and computer-readable storage medium |
| CN112308466A (en) * | 2020-11-26 | 2021-02-02 | 东莞市盟大塑化科技有限公司 | Enterprise qualification auditing method and device, computer equipment and storage medium |
| CN112579579A (en) * | 2019-09-30 | 2021-03-30 | 北京国双科技有限公司 | Material mobile data auditing method and device, storage medium and electronic equipment |
| CN112734352A (en) * | 2019-10-28 | 2021-04-30 | 北京京东尚科信息技术有限公司 | Document auditing method and device based on data dimensionality |
| CN113159175A (en) * | 2021-04-21 | 2021-07-23 | 平安科技（深圳）有限公司 | Data prediction method, device, equipment and storage medium |
| CN113435842A (en) * | 2021-06-28 | 2021-09-24 | 京东科技控股股份有限公司 | Business process processing method and computer equipment |
| CN113705668A (en) * | 2021-08-27 | 2021-11-26 | 创新奇智(广州)科技有限公司 | Method, device, equipment and storage medium for detecting working state of component |
| CN114202399A (en) * | 2021-12-13 | 2022-03-18 | 金蝶软件（中国）有限公司 | Intelligent approval method and related device |
| CN114496196A (en) * | 2022-02-18 | 2022-05-13 | 潍坊医学院附属医院 | Automatic auditing system for clinical biochemical inspection in medical laboratory |
| WO2024021555A1 (en) * | 2022-07-29 | 2024-02-01 | 京东科技信息技术有限公司 | Resource examination and approval method and device, and random forest model training method and device |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8818910B1 (en) * | 2013-11-26 | 2014-08-26 | Comrise, Inc. | Systems and methods for prioritizing job candidates using a decision-tree forest algorithm |
| CN105279691A (en) * | 2014-07-25 | 2016-01-27 | 中国银联股份有限公司 | Financial transaction detection method and equipment based on random forest model |
| CN107092827A (en) * | 2017-03-30 | 2017-08-25 | 中国民航大学 | A kind of Android malware detection method based on improvement forest algorithm |
| CN107766883A (en) * | 2017-10-13 | 2018-03-06 | 华中师范大学 | A kind of optimization random forest classification method and system based on weighted decision tree |
| CN108510507A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨理工大学 | A kind of 3D vertebra CT image active profile dividing methods of diffusion-weighted random forest |
| CN108665159A (en) * | 2018-05-09 | 2018-10-16 | 深圳壹账通智能科技有限公司 | A kind of methods of risk assessment, device, terminal device and storage medium |
| CN109145965A (en) * | 2018-08-02 | 2019-01-04 | 深圳辉煌耀强科技有限公司 | Cell recognition method and device based on random forest disaggregated model |
| CN109214914A (en) * | 2018-08-24 | 2019-01-15 | 厦门集微科技有限公司 | A kind of loan information checking method and device based on communication open platform |
| CN109359669A (en) * | 2018-09-10 | 2019-02-19 | 平安科技（深圳）有限公司 | Method for detecting abnormality, device, computer equipment and storage medium are submitted an expense account in medical insurance |
| CN109389490A (en) * | 2018-09-26 | 2019-02-26 | 深圳壹账通智能科技有限公司 | Loan product matching process, device, computer equipment and storage medium |
| CN109460872A (en) * | 2018-11-14 | 2019-03-12 | 重庆邮电大学 | One kind being lost unbalanced data prediction technique towards mobile communication subscriber |
| CN109564677A (en) * | 2018-11-09 | 2019-04-02 | 香港应用科技研究院有限公司 | Super-resolution synthesis system and method based on random forest grader weighted results |
| CN109726826A (en) * | 2018-12-19 | 2019-05-07 | 东软集团股份有限公司 | Training method, device, storage medium and the electronic equipment of random forest |
| CN109829471A (en) * | 2018-12-19 | 2019-05-31 | 东软集团股份有限公司 | Training method, device, storage medium and the electronic equipment of random forest |
-
2019
- 2019-06-19 CN CN201910533825.2A patent/CN110264342B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8818910B1 (en) * | 2013-11-26 | 2014-08-26 | Comrise, Inc. | Systems and methods for prioritizing job candidates using a decision-tree forest algorithm |
| CN105279691A (en) * | 2014-07-25 | 2016-01-27 | 中国银联股份有限公司 | Financial transaction detection method and equipment based on random forest model |
| CN107092827A (en) * | 2017-03-30 | 2017-08-25 | 中国民航大学 | A kind of Android malware detection method based on improvement forest algorithm |
| CN107766883A (en) * | 2017-10-13 | 2018-03-06 | 华中师范大学 | A kind of optimization random forest classification method and system based on weighted decision tree |
| CN108510507A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨理工大学 | A kind of 3D vertebra CT image active profile dividing methods of diffusion-weighted random forest |
| CN108665159A (en) * | 2018-05-09 | 2018-10-16 | 深圳壹账通智能科技有限公司 | A kind of methods of risk assessment, device, terminal device and storage medium |
| CN109145965A (en) * | 2018-08-02 | 2019-01-04 | 深圳辉煌耀强科技有限公司 | Cell recognition method and device based on random forest disaggregated model |
| CN109214914A (en) * | 2018-08-24 | 2019-01-15 | 厦门集微科技有限公司 | A kind of loan information checking method and device based on communication open platform |
| CN109359669A (en) * | 2018-09-10 | 2019-02-19 | 平安科技（深圳）有限公司 | Method for detecting abnormality, device, computer equipment and storage medium are submitted an expense account in medical insurance |
| CN109389490A (en) * | 2018-09-26 | 2019-02-26 | 深圳壹账通智能科技有限公司 | Loan product matching process, device, computer equipment and storage medium |
| CN109564677A (en) * | 2018-11-09 | 2019-04-02 | 香港应用科技研究院有限公司 | Super-resolution synthesis system and method based on random forest grader weighted results |
| CN109460872A (en) * | 2018-11-14 | 2019-03-12 | 重庆邮电大学 | One kind being lost unbalanced data prediction technique towards mobile communication subscriber |
| CN109726826A (en) * | 2018-12-19 | 2019-05-07 | 东软集团股份有限公司 | Training method, device, storage medium and the electronic equipment of random forest |
| CN109829471A (en) * | 2018-12-19 | 2019-05-31 | 东软集团股份有限公司 | Training method, device, storage medium and the electronic equipment of random forest |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112579579A (en) * | 2019-09-30 | 2021-03-30 | 北京国双科技有限公司 | Material mobile data auditing method and device, storage medium and electronic equipment |
| CN112734352A (en) * | 2019-10-28 | 2021-04-30 | 北京京东尚科信息技术有限公司 | Document auditing method and device based on data dimensionality |
| CN111612166B (en) * | 2020-04-23 | 2022-10-25 | 中国科学院计算机网络信息中心 | Reimbursement time prediction method based on machine learning |
| CN111612166A (en) * | 2020-04-23 | 2020-09-01 | 中国科学院计算机网络信息中心 | Reimbursement time prediction method based on machine learning |
| CN112258135A (en) * | 2020-05-15 | 2021-01-22 | 北京沃东天骏信息技术有限公司 | Method and device for auditing prescription data and computer-readable storage medium |
| CN112308466A (en) * | 2020-11-26 | 2021-02-02 | 东莞市盟大塑化科技有限公司 | Enterprise qualification auditing method and device, computer equipment and storage medium |
| CN113159175A (en) * | 2021-04-21 | 2021-07-23 | 平安科技（深圳）有限公司 | Data prediction method, device, equipment and storage medium |
| CN113159175B (en) * | 2021-04-21 | 2023-06-06 | 平安科技（深圳）有限公司 | Data prediction method, device, equipment and storage medium |
| CN113435842A (en) * | 2021-06-28 | 2021-09-24 | 京东科技控股股份有限公司 | Business process processing method and computer equipment |
| CN113705668A (en) * | 2021-08-27 | 2021-11-26 | 创新奇智(广州)科技有限公司 | Method, device, equipment and storage medium for detecting working state of component |
| CN114202399A (en) * | 2021-12-13 | 2022-03-18 | 金蝶软件（中国）有限公司 | Intelligent approval method and related device |
| CN114496196A (en) * | 2022-02-18 | 2022-05-13 | 潍坊医学院附属医院 | Automatic auditing system for clinical biochemical inspection in medical laboratory |
| WO2024021555A1 (en) * | 2022-07-29 | 2024-02-01 | 京东科技信息技术有限公司 | Resource examination and approval method and device, and random forest model training method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110264342B (en) | 2024-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110264342A (en) | A kind of business audit method and device based on machine learning | |
| Guo et al. | Blockchain application and outlook in the banking industry | |
| CN110458693A (en) | A kind of automatic measures and procedures for the examination and approval of business loan, device, storage medium and electronic equipment | |
| CN110276620A (en) | It is a kind of to determine the method and device traded extremely | |
| CN108665366A (en) | Determine method, terminal device and the computer readable storage medium of consumer's risk grade | |
| CN108230140A (en) | The method and apparatus of pushed information, the method and apparatus for determining input default value | |
| CN109740792A (en) | Data predication method, system, terminal and computer storage medium | |
| CN108629685A (en) | loan product attribute determining method and server | |
| Vasylieva et al. | Scientific and methodical approaches to determining the center-orientation of financial conglomerates with the factor and cluster analysis | |
| Pellegrino et al. | Digital financing for SMEs' recovery in the post-COVID era: A bibliometric review | |
| CN102915466A (en) | Data processing method and device | |
| US11798100B2 (en) | Transaction counterpart identification | |
| Ramzan et al. | Impact of asset preferences on firm performance over its life cycle: Is agency theory or neo‐classical theory more relevant? | |
| Biswas et al. | Automated credit assessment framework using ETL process and machine learning | |
| CN110276692B (en) | Method and device for processing transaction data | |
| Kaminskyi et al. | Information technology model for customer relationship management of nonbank lenders: coupling profitability and risk | |
| Bakoeva | The Essence Of Transformation Of The Banking Sector In The Context Of The Implementation Of Modern Financial Technologies | |
| Vovchak et al. | Bank business modeling and levels of non-performing loans: Perspectives of international risk factors in Ukraine | |
| Reddy et al. | Peer to peer lending, default prediction-evidence from lending club | |
| Egorova et al. | Digital currency in the development of payment systems on the bitcoin platform | |
| Topaloglou et al. | Investors’ Behavior in Cryptocurrency Market | |
| Reyes-Mercado et al. | How fintech was born? | |
| Liu et al. | Research on the identification and prediction of default risk of online lending platform customers | |
| CN109657897A (en) | Mobility notch method of adjustment, device, equipment and storage medium | |
| Chen et al. | Risk and Return of Securities Investment in the Context of Big Data and Artificial Intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |