CN110232525A - A kind of business risk monitoring method, device, server and storage medium - Google Patents
A kind of business risk monitoring method, device, server and storage medium Download PDFInfo
- Publication number
- CN110232525A CN110232525A CN201910517033.6A CN201910517033A CN110232525A CN 110232525 A CN110232525 A CN 110232525A CN 201910517033 A CN201910517033 A CN 201910517033A CN 110232525 A CN110232525 A CN 110232525A
- Authority
- CN
- China
- Prior art keywords
- enterprise
- dimension
- risk
- characteristic
- monitoring
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 141
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000009826 distribution Methods 0.000 claims description 10
- 238000012806 monitoring device Methods 0.000 claims description 7
- 238000004519 manufacturing process Methods 0.000 claims description 6
- 230000005055 memory storage Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 8
- 238000010586 diagram Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 238000012937 correction Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
Abstract
The application provides a kind of business risk monitoring method, device, server and storage medium, this method is by determining at least one enterprise of pending monitoring and the risk classifications of the pending monitoring of at least one enterprise, and it is directed to and each dimension at least one matched dimension of the risk classifications, each enterprise is obtained at least one enterprise respectively in the characteristic of the dimension, and then combine the characteristic of each enterprise each dimension at least one dimension at least one enterprise to the mode of at least one enterprise progress risk monitoring and control based on isolated forest algorithm, available risk monitoring and control result is to realize the monitoring to business risk.The application is not needed by expert's artificial constructed risk monitoring and control rule, thus can to avoid existing enterprise's risk monitoring and control technology because needing the artificial constructed risk monitoring and control rule of expert, caused by time-consuming and laborious problem.
Description
Technical field
The present invention relates to risk monitoring and control technical field, more specifically to a kind of business risk monitoring method, device,
Server and storage medium.
Background technique
Business risk monitoring occupies very important effect in risk monitoring and control field, not only can find that in time there are wind
The enterprise of danger guarantees the good development of enterprise, but also can make prevention before risk generation in order to enterprise, guarantees enterprise
Safety.
Business risk monitoring at present is dependent on risk monitoring and control expert according to the knowledge man of financial accounting, business administration etc.
Work constructs business risk monitoring rules, acquires data disclosed in enterprise, the every risk specified by enterprise diagnosis monitoring rules
Monitor control index realizes the monitoring to business risk.Although the monitoring to business risk may be implemented in this method, but because needing expert
Artificial constructed risk monitoring and control rule, therefore usually there is a problem of time-consuming and laborious.
In view of this, a kind of business risk monitoring method, device, server and storage medium how are provided, to avoid existing
The problem for having business risk monitoring technology time-consuming and laborious, is a problem to be solved.
Summary of the invention
In view of this, to solve the above problems, the present invention provide a kind of business risk monitoring method, device, server and
Storage medium, to avoid the time-consuming and laborious problem of existing enterprise's risk monitoring and control technology.Technical solution is as follows:
A kind of business risk monitoring method, comprising:
Determine at least one enterprise of pending monitoring and the risk classifications of the pending monitoring of at least one enterprise;
The enterprise is obtained in the feature with each of at least one matched dimension of risk classifications dimension
Data;
Based on isolated forest algorithm in conjunction at least one described enterprise each dimension at least one described dimension
Characteristic, at least one described enterprise carry out risk monitoring and control, obtain risk monitoring and control result.
A kind of business risk monitoring device, comprising:
Information determination unit, at least one enterprise and at least one described enterprise for determining pending monitoring wait for into
The risk classifications of row monitoring;
Characteristic acquiring unit, for obtain the enterprise at least one matched dimension of the risk classifications
Each of the dimension characteristic;
Risk monitoring and control unit, for being based on isolated forest algorithm in conjunction at least one described enterprise at least one described dimension
The characteristic of each dimension in degree carries out risk monitoring and control at least one described enterprise, obtains risk monitoring and control result.
A kind of server, comprising: at least one processor and at least one processor;The memory is stored with program,
The processor calls the program of the memory storage, and described program is for realizing the business risk monitoring method.
A kind of storage medium is stored with computer executable instructions in the storage medium, and the computer is executable to be referred to
It enables for executing the business risk monitoring method.
The application provides a kind of business risk monitoring method, device, server and storage medium, this method by determine to
The risk classifications of at least one enterprise being monitored and the pending monitoring of at least one enterprise, and be directed to and the risk class
Each dimension at least one matched dimension of type obtains at least one enterprise each enterprise in the feature of the dimension respectively
Data, and then each enterprise each dimension at least one dimension at least one enterprise is combined based on isolated forest algorithm
Characteristic carries out the mode of risk monitoring and control at least one enterprise, and available risk monitoring and control result is to realize to business risk
Monitoring.
The application does not need during realizing business risk monitoring by expert's artificial constructed risk monitoring and control rule, because
This can to avoid existing enterprise's risk monitoring and control technology because needing expert's artificial constructed risk monitoring and control rule, caused by it is time-consuming and laborious
Problem.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of hardware block diagram of server provided by the embodiments of the present application;
Fig. 2 is a kind of business risk monitoring method flow chart provided by the embodiments of the present application;
Fig. 3 is another business risk monitoring method flow chart provided by the embodiments of the present application;
Fig. 4 be it is provided by the embodiments of the present application it is a kind of according to enterprise each at least one enterprise dimension characteristic
According to determining the method flow diagram of the cut-point of dimension;
Fig. 5 be the cut-point using dimension in a kind of business risk monitoring method provided by the embodiments of the present application correct to
Method flow diagram of each enterprise in the characteristic of the dimension in a few enterprise;
Fig. 6 is that one kind provided by the embodiments of the present application is based on isolated forest algorithm in conjunction at least one enterprise at least one
The revised characteristic of each dimension in dimension carries out risk monitoring and control at least one enterprise, obtains risk monitoring and control result
Method flow diagram;
The method flow for the reason of Fig. 7 is a kind of risk of determining enterprise's occurrence risk type provided by the embodiments of the present application
Figure;
Fig. 8 is a kind of structural schematic diagram of business risk monitoring device provided by the embodiments of the present application.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Embodiment:
The embodiment of the present application provides a kind of business risk monitoring method, and this method is independent of the wind artificial constructed by expert
Dangerous monitoring rules (risk monitoring and control rule includes multinomial risk monitoring and control index), it can thus be avoided existing enterprise's risk monitoring and control skill
Art is realized with the risk monitoring and control index indicated based on risk monitoring and control rule to enterprise because needing the artificial constructed risk monitoring and control rule of expert
The monitoring of risk, existing time-consuming and laborious problem.
A kind of business risk monitoring method provided by the embodiments of the present application can be applied to server, which can be net
Network side provides the service equipment of service for user, may be the server cluster of multiple servers composition, it is also possible to separate unit
Server.
Optionally, Fig. 1 shows the hardware block diagram of server, and referring to Fig.1, the hardware configuration of server can wrap
It includes: processor 11, communication interface 12, memory 13 and communication bus 14;
In embodiments of the present invention, processor 11, communication interface 12, memory 13, communication bus 14 quantity can be with
For at least one, and processor 11, communication interface 12, memory 13 complete mutual communication by communication bus 14;
Processor 11 may be a central processor CPU or specific integrated circuit ASIC (Application
Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention
Road etc.;
Memory 13 may include high speed RAM memory, it is also possible to further include nonvolatile memory (non-volatile
Memory) etc., a for example, at least magnetic disk storage;
Wherein, memory is stored with program, the program that processor can call memory to store, and program is used for:
Determine the risk classifications of the pending monitoring of at least one enterprise and at least one enterprise of pending monitoring;
Enterprise is obtained in the characteristic with each dimension at least one matched dimension of risk classifications;
The characteristic of at least one enterprise each dimension at least one dimension is combined based on isolated forest algorithm, it is right
At least one enterprise carries out risk monitoring and control, obtains risk monitoring and control result.
For the ease of the understanding to the business risk monitoring method for being suitable for above-mentioned server, now the embodiment of the present application is mentioned
A kind of business risk monitoring method supplied describes in detail.
Fig. 2 is a kind of business risk monitoring method flow chart provided by the embodiments of the present application.
As shown in Fig. 2, this method comprises:
S201, the pending monitoring of at least one enterprise for determining pending monitoring and at least one enterprise risk class
Type;
In the embodiment of the present application, risk classifications include financial risk type, Industry risk type, real example risk classifications and
It is any one or more in negative public sentiment risk classifications.
For each risk classifications, can preset and at least one matched dimension of the risk classifications;For example, in advance
Be arranged at least one dimension of financial risk type matching, preset with Industry risk type matching at least one dimension
Degree, preset at least one matched dimension of real example risk classifications, preset matched with negative public sentiment risk classifications
At least one dimension.
Different at least one matched dimensions of risk classifications institute is different, for example, at least with financial risk type matching
One dimension is different from least one dimension with Industry risk type matching.
In the embodiment of the present application, it is preferred that it is big to be segmented into three at least one dimension of financial risk type matching
Class, these three types are respectively as follows: general profit of each season, general cash stream, purpose balance and are in debt.
Wherein, a season general profit can regard a dimension as, and general cash stream can regard a dimension as,
Purpose balance debt can regard a dimension as.For example, an enterprise is at least one dimension with financial risk type matching
The characteristic of degree may include that the general profit of the first quarter is 689187341.1 yuan.
In the embodiment of the present application, it is preferred that include industrial and commercial punishment at least one matched dimension of real example risk classifications
Dimension, administrative penalty dimension etc..For example, an enterprise is in the characteristic at least one matched dimension of real example risk classifications
According to may include: 20181005 order stop publication Unlawful cost, impose a fine 18778.9 yuan.
In the embodiment of the present application, it is preferred that Shen ten thousand can be referred to at least one dimension of Industry risk type matching
Classification system, construct three-level trade classification, be subdivided into 230 industries.For example, an enterprise with Industry risk type matching
At least one dimension characteristic include take trade classification calculate industrial characteristic: XX industry net profit margin 8.3%.
In the embodiment of the present application, it is preferred that at least one matched dimension of negative public sentiment risk classifications include enterprise
The negative public sentiment frequency of nearest period.For example, an enterprise with negative at least one matched dimension of public sentiment risk classifications
Characteristic may include: the negative public sentiment of appearance of enterprise 10 times, this 10 times negative public sentiment can be divided into 3 grades, respectively
For high grade, advanced tiers and intermediate grade have in this 10 times negative public sentiments 3 negative public sentiments to belong to high grade, and 5 times
Negative public sentiment belongs to advanced tiers, and 2 times negative public sentiment belongs to intermediate grade;Belong in the negative public sentiment of 10 times of appearance of enterprise high
The ratio of the negative public sentiment number of grade, the negative public sentiment number for belonging to advanced tiers and the negative public sentiment number for belonging to intermediate grade
Example is 3:5:2.
It is above only provided by the embodiments of the present application preferred from different at least one matched dimension of risk classifications
Content, in relation to risk classifications at least one matched dimension particular content, inventor can set according to their own needs
It sets, it is not limited here.
S202, enterprise is obtained in the characteristic with each dimension at least one matched dimension of risk classifications;
In the embodiment of the present application, the risk classifications and the pending risk classifications of pending monitoring can be determined
At least one enterprise of monitoring;And then determine pre-set and at least one matched dimension of the risk classifications, and for extremely
Each dimension in a few dimension obtains at least one enterprise each enterprise in the characteristic of the dimension.
For example, enterprise is enterprise first in the characteristic of the dimension when dimension is the general profit of the first quarter
The general profit in season.
In the embodiment of the present application, for a dimension, when obtaining characteristic of some enterprise in the dimension, if hair
Now the enterprise is not disclosed in the characteristic of dimension, can set infinitely great in the characteristic of the dimension for the enterprise, that is,
By it is infinitely great as the enterprise the dimension characteristic.
S203, the characteristic that at least one enterprise each dimension at least one dimension is combined based on isolated forest algorithm
According to, at least one enterprise carry out risk monitoring and control, obtain risk monitoring and control result.
In the embodiment of the present application, for each dimension at least one matched dimension of risk classifications,
It determines that each enterprise is after the characteristic of the dimension at least one enterprise, isolated forest algorithm can be run, it will at least one
Each enterprise characteristic of each dimension at least one dimension is input to the root node of isolated forest algorithm in a enterprise,
The characteristic of each enterprise each dimension at least one dimension at least one enterprise is carried out by isolating forest algorithm
Segmentation obtains the risk monitoring and control result of the risk classifications to realize to the risk monitoring and control of at least one enterprise.
In the embodiment of the present application, when risk classifications are financial risk type, in order to further increase to financial risk
The accuracy of the risk monitoring and control result of type, the embodiment of the present application also provides another business risk monitoring method, referring to Fig. 3.
As shown in figure 3, this method comprises:
S301, the pending monitoring of at least one enterprise for determining pending monitoring and at least one enterprise risk class
Type;
In the embodiment of the present application, at least one enterprise and at least one enterprise of pending monitoring can be determined
The risk classifications of pending monitoring, at this point, the risk classifications of the pending monitoring of at least one enterprise are financial risk type.
S302, enterprise is obtained in the characteristic with each dimension at least one matched dimension of risk classifications;
In the embodiment of the present application, determining at least one dimension with financial risk type matching, for it is identified extremely
Each dimension in a few dimension, determines that each enterprise is in the characteristic of the dimension at least one enterprise.
For example, if with including two dimensions, the two dimensions difference at least one dimension of financial risk type matching
For " the general profit of the first quarter " dimension and " the general profit of the second quarter " dimension;If at least one enterprise includes enterprise
1, enterprise 2 and enterprise 3, then it needs to be determined that enterprise 1 exists in the characteristic of " the general profit of the first quarter " dimension and enterprise 1
The characteristic of " the general profit of the second quarter " dimension, characteristic of the enterprise 2 in " the general profit of the first quarter " dimension
And enterprise 2 is in the characteristic of " the general profit of the second quarter " dimension, and, enterprise 3 is in " the general benefit of the first quarter
Profit " dimension characteristic and enterprise 3 " the general profit of the second quarter " dimension characteristic.
S303, for each dimension at least one dimension, according to enterprise each at least one enterprise in the dimension
Characteristic, determine the cut-point of the dimension;
In the embodiment of the present application, each at least one enterprise for each dimension at least one dimension
Enterprise, can be with based on each enterprise at least one enterprise in the characteristic of the dimension in the dimension all existing characteristics data
Determine the cut-point of the dimension.
In the embodiment of the present application, which is determined in the characteristic of the dimension according to enterprise each at least one enterprise
The mode of the cut-point of degree may refer to hereafter to " true in the characteristic of dimension according to enterprise each at least one enterprise
Determine the cut-point of dimension " detailed description of step, this will not be repeated here.
S304, each enterprise is corrected at least one enterprise in the characteristic of dimension using the cut-point of dimension;
In the embodiment of the present application, each at least one enterprise for each dimension at least one dimension
All there is the characteristic of the dimension in enterprise, after determining the cut-point of the dimension, can be distinguished based on the cut-point of the dimension
Each enterprise at least one enterprise is modified in the characteristic of the dimension, obtains revised characteristic.
For example, can according to enterprise 1 in the characteristic of " the general profit of the first quarter " dimension, enterprise 2 in the " first season
The characteristic of the general profit of degree " dimension and enterprise 3 calculate in the characteristic of " the general profit of the first quarter " dimension
The cut-point of " the general profit of the first quarter " dimension;And then the cut-point based on " the general profit of the first quarter " dimension is corrected
Enterprise 1 obtains enterprise 1 in " the general profit of the first quarter " dimension in the characteristic of " the general profit of the first quarter " dimension
Revised characteristic, based on " the general profit of the first quarter " dimension cut-point amendment enterprise 2 " first quarter
The characteristic of general profit " dimension obtains enterprise 2 in the revised characteristic of " the general profit of the first quarter " dimension,
And the cut-point amendment enterprise 3 based on " the general profit of the first quarter " dimension is in " the general profit of the first quarter " dimension
Characteristic obtains enterprise 3 in the revised characteristic of " the general profit of the first quarter " dimension.
S305, combine at least one enterprise at least one dimension after the amendment of each dimension based on isolated forest algorithm
Characteristic, at least one enterprise carry out risk monitoring and control, obtain risk monitoring and control result.
In the embodiment of the present application, for each dimension at least one matched dimension of risk classifications,
Determine that each enterprise can run isolated forest algorithm after the revised characteristic of the dimension at least one enterprise,
By enterprise each at least one enterprise at least one dimension the revised characteristic of each dimension be input to it is isolated
The root node of forest algorithm, by isolated forest algorithm to each enterprise at least one enterprise each dimension at least one dimension
The revised characteristic of degree is split, and to realize to the risk monitoring and control of at least one enterprise, obtains the risk classifications
Risk monitoring and control result.
In order to make it easy to understand, existing to " for a dimension, according to enterprise each at least one enterprise in the dimension
Characteristic determine the cut-point of the dimension " mode describes in detail.
In the embodiment of the present application, there are a target dimensions at least one dimension, are financial risk in risk classifications
It can be " net profit " dimension with the target dimension at least one dimension of financial risk type matching when type.Above only
It is only the preferred content of target dimension provided by the embodiments of the present application, the particular content in relation to target dimension, inventor can basis
The demand of oneself is configured, it is not limited here.
For each dimension at least one dimension with financial risk type matching, need to calculate the dimension
Cut-point.Referring to fig. 4 for it is provided by the embodiments of the present application it is a kind of according to enterprise each at least one enterprise dimension feature
Data determine the method flow diagram of the cut-point of dimension.
As shown in figure 4, this method comprises:
S401, it counts at least one enterprise in enterprise of the characteristic of dimension greater than 0, obtains the first number of the enterprise;
S402, it counts at least one enterprise in enterprise of the characteristic of dimension less than 0, obtains the second number of the enterprise;
In the embodiment of the present application, it is directed to a dimension, determines that the mode of the cut-point of the dimension can be with are as follows: determine extremely
Each enterprise and counts the characteristic at least one enterprise in the dimension in the characteristic of the dimension in a few enterprise
The quantity of enterprise greater than 0, and using the quantity as the first number of the enterprise;And it counts at least one enterprise in the dimension
The quantity of enterprise of the characteristic less than 0, and using the quantity as the second number of the enterprise.
S403, judge whether dimension is target dimension;If dimension is target dimension, step S404 is executed;If dimension is not
Target dimension executes step S405;
S404, it is based on the first number of the enterprise and the second number of the enterprise, calculates the cut-point of dimension.
In the embodiment of the present application, it is directed to a dimension, if the dimension is not target dimension, counts at least one enterprise
In the dimension in the first number of the enterprise of enterprise of the characteristic of the dimension greater than 0 and at least one enterprise in industry
After second number of the enterprise of enterprise of the characteristic less than 0, the first number of the enterprise and the second number of the enterprise can be summed up
It calculates, obtains number of the enterprise;And then the first number of the enterprise is occupied to the ratio of number of the enterprise, the cut-point as the dimension.
For example, a dimension is directed to, if the dimension is not target dimension, if at this at least one enterprise counted
First number of the enterprise of enterprise of the characteristic of dimension greater than 0 is the feature in 90 and at least one enterprise in the dimension
After second number of the enterprise of enterprise of the data less than 0 is 219,309 (90+219=309) can be regard as number of the enterprise, in turn
Cut-point by 0.29 (90/309=0.29) [note that being rounded up] herein as the dimension.
S405, the prior distribution based on the first number of the enterprise and the second number of the enterprise amendment target dimension, obtain dimension
Cut-point.
In the embodiment of the present application, the mode for calculating the prior distribution of target dimension can be with are as follows: counts at least one enterprise
In target dimension characteristic greater than 0 enterprise quantity, obtain third number of the enterprise;And at least one enterprise
The quantity of enterprise of the characteristic of target dimension less than 0, obtains the 4th number of the enterprise, by third number of the enterprise and the 4th enterprise
Quantity brings prior distribution formula into, obtains the prior distribution of target dimension.
It further,, can be according at least one enterprise if the dimension is not target dimension for a dimension
Characteristic in first number of the enterprise of the characteristic of the dimension greater than 0 and at least one enterprise in the dimension is less than 0
The second number of the enterprise, the prior distribution of target dimension is modified, to obtain the cut-point of the dimension.
As a kind of preferred embodiment of the embodiment of the present application, for a dimension, if the dimension is not target
Dimension can count at least one enterprise in first number of the enterprise of the characteristic of the dimension greater than 0 and at least one enterprise
In second number of the enterprise of the dimensional characteristics data less than 0 in industry, the priori point of target dimension is corrected by the first number of the enterprise
Third number of the enterprise in cloth, and pass through the second number of the enterprise correct target dimension prior distribution in the 4th enterprise's number
Amount, to obtain revised distribution (the revised distribution may be considered Posterior distrbutionp, that is, the cut-point of the dimension).
In the embodiment of the present application, for a dimension, enterprise is equal to 0 in the characteristic of the dimension sometimes, this
When, it can be set and the statistics of above-mentioned number of the enterprise is not impacted, it also can be set, the statistics of number of the enterprise is caused
It influences;For a dimension, the enterprise that a characteristic in the dimension is equal to 0 can be regarded as one in the dimension
Enterprise of the characteristic of degree greater than 0 carries out number of the enterprise statistics, a characteristic in the dimension can also be equal to 0
Enterprise regards enterprise of the characteristic in the dimension less than 0 as and carries out number of the enterprise statistics.
In order to make it easy to understand, existing to utilizing dimension in a kind of business risk monitoring method provided by the embodiments of the present application
Cut-point is corrected each enterprise at least one enterprise and is described in detail in the method for the characteristic of the dimension, please specifically join
See Fig. 5.
As shown in figure 5, this method comprises:
Whether S501, detection enterprise are greater than the cut-point of dimension in the characteristic of dimension;If enterprise is in the feature of dimension
Data are not more than the cut-point of dimension, execute step S502;If enterprise is greater than the cut-point of dimension in the characteristic of dimension, hold
Row step S503;
S502, retain enterprise in the characteristic of dimension;
S503, enterprise is set to preset value in the characteristic of dimension;
In the embodiment of the present application, after determining the cut-point of dimension, if based on the cut-point of the dimension to enterprise at this
The mode that the characteristic of dimension is modified are as follows: judge whether enterprise is greater than the segmentation of the dimension in the characteristic of the dimension
Point, if enterprise the dimension characteristic be not more than the dimension cut-point, not to the enterprise the dimension characteristic
According to being modified, i.e., reservation enterprise the dimension characteristic (for example, the cut-point of dimension be 0.29 when, if enterprise is at this
The characteristic of dimension is 0.27, then is not modified to the enterprise in the characteristic of the dimension, i.e., the enterprise is in the dimension
Characteristic be still 0.27);If enterprise is greater than the cut-point of the dimension in the characteristic of the dimension, by the enterprise at this
The characteristic of dimension be set to preset value (cut-point of dimension be 0.29 when, if enterprise the dimension characteristic be 0.3,
The enterprise is then set to preset value in the characteristic of the dimension).
In the embodiment of the present application, preset value can be 1,20 etc., and above is only provided by the embodiments of the present application pre-
If the preferred content of value, the specific value inventor in relation to preset value can be configured according to their own needs, not limit herein
It is fixed.
For example, if, can be with the cut-point of dimension for 0.29, enterprise be in the characteristic of the dimension when preset value is 1
0.3, then the enterprise can be set to 1 in the characteristic of the dimension.
S504, each enterprise at least one enterprise is normalized in the characteristic of dimension, is obtained at least
Target signature data of each enterprise in dimension in one enterprise.
In the embodiment of the present application, for a dimension, in the cut-point for determining the dimension, based on cut-point at least one
Each enterprise is after the characteristic of the dimension is performed both by a step S501-S503 in a enterprise, it is available at least one
Characteristic of each enterprise after the initial correction of the dimension in enterprise;And then it can also be further at least one enterprise
Characteristic of each enterprise after the initial correction of the dimension is normalized, and obtains each enterprise at least one enterprise
Revised characteristic of the industry in the dimension, that is, target signature data of each enterprise in the dimension at least one enterprise.
One kind as the embodiment of the present application is preferably implementation, to an enterprise after the initial correction of a dimension
The mode that is normalized of characteristic can be with are as follows: obtain at least one enterprise the characteristic of the dimension most
Big value and minimum value;It calculates characteristic of the enterprise after the initial correction of the dimension and subtracts the result of minimum value (in order to just
In differentiation, the result is temporarily known as the first data);Calculate the maximum value subtract minimum value result (for the ease of distinguish, temporarily will
The result is known as the second data);Using the first data divided by the second data result as the enterprise the dimension target signature
Data.
In the embodiment of the present application, each enterprise at least one enterprise is being obtained in each of at least one dimension
It, can be using obtained all target signature data as input information input to isolated forest after the target signature data of dimension
The root node of algorithm, and isolated forest algorithm is run, target signature data are split based on isolated forest algorithm, realization pair
The risk monitoring and control of at least one enterprise obtains risk monitoring and control result.
It can be to each enterprise at least one enterprise to each dimension at least one dimension when isolated forest algorithm operation
The target signature data of degree carry out random division, to generate random binary tree, and for each enterprise at least one enterprise
For industry, which can fall into a leaf node in the random binary tree.The embodiment of the present application is in order to guarantee to enterprise's wind
The accuracy nearly monitored can run repeatedly isolated forest algorithm, to generate more random binary trees.
In order to make it easy to understand, existing be based on isolated forest algorithm in conjunction at least one enterprise to one of the embodiment of the present application
The revised characteristic of each dimension at least one dimension carries out risk monitoring and control at least one enterprise, obtains wind
The method of dangerous monitored results is described in detail, and specifically refers to Fig. 6.
As shown in fig. 6, this method comprises:
S601, based on isolated forest algorithm at least one enterprise at least one dimension each dimension it is revised
Characteristic is split, and obtains random binary tree;
In the embodiment of the present application, repeatedly isolated forest algorithm will at least when each run isolates forest algorithm for operation
Each enterprise is to the target signature data of each dimension at least one dimension as the defeated of isolated forest algorithm in one enterprise
Enter information, input information is split by isolated forest algorithm, obtains a random binary tree.
Correspondingly, once isolated forest algorithm obtains a random binary tree for every operation, isolated forest algorithm is run repeatedly
More random binary trees are obtained, and then are realized based on more obtained random binary trees to the business risk of at least one enterprise
Monitoring.
S602, for each enterprise at least one enterprise, in multiple isolated obtained more of the forest algorithm of operation
In random binary tree, the leaf node that enterprise falls into every random binary tree is counted;
In the embodiment of the present application, once isolated forest algorithm obtains a random binary tree for every operation, runs repeatedly lonely
Vertical forest algorithm obtains more random binary trees;For each enterprise at least one enterprise, determine the enterprise respectively every
The leaf node fallen into random binary tree, and then the leaf node fallen into each random binary tree based on the enterprise,
Determine the enterprise with the presence or absence of risk.
S603, the leaf node fallen into every random binary tree according to enterprise determine enterprise with the presence or absence of risk class
The risk of type.
In the embodiment of the present application, for every random binary tree in more random binary trees, determine that enterprise falls
Enter depth of the leaf node of the random binary tree of this in the random binary tree of this;And then it is directed to any one depth, statistics
The number of the leaf node for the depth that the enterprise falls into more random binary trees is obtained, and then is obtained and risk classifications
The risk conditions matched, judge whether the number for the leaf node that enterprise falls into predetermined depth in more random binary trees meets this
Risk conditions, if satisfied, determining enterprise, there are the risks of the risk classifications, if not satisfied, determining that the risk class is not present in enterprise
The risk of type.
For example, 5 isolated forest algorithms of operation, obtain 5 random binary trees, respectively random binary tree 1, random y-bend
Tree 2, random binary tree 3, random binary tree 4 and random binary tree 5, enterprise 1 fall into the leaf node of random binary tree 1 random
Depth in binary tree 1 is 2;It is 3 that enterprise 1, which falls into depth of the leaf node of random binary tree 2 in random binary tree 2,;Enterprise
It is 2 that industry 1, which falls into depth of the leaf node of random binary tree 3 in random binary tree 3,;Enterprise 1 falls into the leaf of random binary tree 4
Depth of the child node in random binary tree 4 is 2;Enterprise 1 falls into the leaf node of random binary tree 5 in random binary tree 5
Depth be 3;Then statistics obtains enterprise 1 to fall into the number for the leaf node that depth is 2 being 3, and enterprise 1 falls into the leaf that depth is 3
The number of child node is 2.
As a kind of preferred embodiment of the embodiment of the present application, it is default to judge that enterprise falls into more random binary trees
Whether the number of the leaf node of depth meets the preset modes with the matched risk conditions of risk classifications can be with are as follows: judgement enterprise
Whether the number for the leaf node that industry falls into predetermined depth in more random binary trees is greater than threshold value, if more than, it is determined that enterprise
There are the risks of risk classifications for industry;If being not more than, it is determined that the risk of risk classifications is not present in enterprise.
Wherein, predetermined depth can be made of at least one depth, for example, predetermined depth includes depth 1-3, predetermined depth
It include depth 3 etc. including depth 2, predetermined depth.It is above only the preferred side of predetermined depth provided by the embodiments of the present application
Formula, the particular content in relation to predetermined depth, inventor can be configured according to their own needs, it is not limited here.
For example, predetermined depth is depth 2, when threshold value is 3, however, it is determined that enterprise 1 falls into depth in more random binary trees
Number for 2 leaf node is greater than 3, it is determined that there are the risks of risk classifications for enterprise 1;If it is determined that enterprise 1 is random at more
The number for the leaf node that depth is 2 is fallen into binary tree no more than 3, it is determined that the risk of risk classifications is not present in enterprise 1.
For example, predetermined depth is depth 2-3, when threshold value is 5, however, it is determined that enterprise 1 falls into depth in more random binary trees
The number of the leaf node of degree 2 and enterprise 1 fall into the number for the leaf node that depth is 3 and big in more random binary trees
In 5, it is determined that there are the risks of risk classifications for enterprise 1;If it is determined that it is 2 that enterprise 1 falls into depth in more random binary trees
The sum that the number of leaf node and enterprise 1 fall into the number for the leaf node that depth is 3 in more random binary trees is not more than
5, it is determined that the risk of risk classifications is not present in enterprise 1.
Further, a kind of business risk monitoring method provided by the embodiments of the present application, determining enterprise, there are risk classes
When the risk of type, the reason of can also further determining that enterprise there are the risks of risk classifications.
Determine enterprise whether there is risk classifications risk when, need first to determine with risk classifications it is matched at least one
Dimension, and the characteristic of each dimension of the enterprise at least one dimension is obtained, and carry out respectively to each characteristic
Amendment obtains revised characteristic (that is, target signature data);And then it is determined based on each target signature data of enterprise
Enterprise whether there is the risk of risk classifications.
It, can be with when each target signature data based on enterprise determine risk of the enterprise there are risk classifications based on this
Determination is which target signature data of enterprise result in the risk that risk classifications have occurred in enterprise, and then cause identified
Dimension belonging to the target signature of the risk of risk classifications has occurred in enterprise, is determined as the original of the risk of enterprise's occurrence risk type
Cause.
Isolated forest algorithm, can be each at least one dimension by enterprise each at least one enterprise in operation
The target signature data of dimension are input to the root node of isolated forest algorithm, and the father node in random binary tree randomly chooses one
Dimension to each enterprise of the leaf node for not falling within random binary tree at least one enterprise currently the dimension feature
Data are split.
In conjunction with above-mentioned general character, at this to a kind of risk of determining enterprise's occurrence risk type provided by the embodiments of the present application
The method of reason is described in detail, and specifically refers to Fig. 7.
As shown in fig. 7, this method comprises:
S701, it is directed to every random binary tree, obtains the target signature for the enterprise that random binary tree uses when generating
Destination node belonging to data and target signature data, destination node belonging to target signature data are base in random binary tree
In the node that the affiliated dimension of target signature data is split at least one enterprise in the target signature data of the dimension;
In the embodiment of the present application, however, it is determined that enterprise is there are when the risk of risk classifications, and can determining enterprise, there are the wind
The reason of risk of dangerous type, this method can be with are as follows: for every obtained random binary tree, determine the random binary tree in life
At when the target signature data of the enterprise that use and the target signature data in the random binary tree of this belonging to mesh
Mark node.
Wherein, target signature data destination node affiliated in the random binary tree of this are as follows: in the random binary tree of this
Based on the father node that target signature data of the dimension belonging to the target signature data to the dimension in Target Enterprise are split,
Target Enterprise is each enterprise that the leaf node of the random binary tree of this is fallen at least one enterprise.
S702, each target signature data that enterprise is used when more random binary trees generate and each mesh are counted
Mark characteristic destination node affiliated in every random binary tree respectively;
For example, 3 isolated forest algorithms of operation, obtain 3 random binary trees, respectively random binary tree 6, random y-bend
Tree 7 and random binary tree 8.If determining enterprise 1 based on random binary tree 6, random binary tree 7 and random binary tree 8, there are risks
When the risk of type, each target signature data of the available enterprise 1 used when generating random binary tree 6 and
Target signature data destination node affiliated in random binary tree 6;Obtain the enterprise used when generating random binary tree 7
The each target signature data and target signature data of industry 1 destination node affiliated in random binary tree 7;And it obtains
The each target signature data and target signature data of the enterprise 1 used when generating random binary tree 8 are random two
Destination node belonging in fork tree 8.
Each target signature data of the enterprise 1 used when generating random binary tree 6 can be target signature number
According to 1 and target signature data 2, wherein the destination node affiliated in random binary tree 6 of target signature data 1 is node 1, mesh
Marking the destination node affiliated in random binary tree 6 of characteristic 2 is node 2.
Each target signature data of the enterprise 1 used when generating random binary tree 7 can be target signature number
According to 1 and target signature data 3, wherein the destination node affiliated in random binary tree 7 of target signature data 1 is node 3, mesh
Marking the destination node affiliated in random binary tree 7 of characteristic 3 is node 4.
Each target signature data of the enterprise 1 used when generating random binary tree 8 can be target signature number
According to 1 and target signature data 2, wherein the destination node affiliated in random binary tree 8 of target signature data 1 is node 5, mesh
Marking the destination node affiliated in random binary tree 8 of characteristic 2 is node 6.
Based on this, it can determine that the target signature data 1 of enterprise 1 destination node affiliated in more random binary trees is
Node 1 in respectively random binary tree 6, the node 3 in random binary tree 7 and the node 5 in random binary tree 8;Enterprise 1
Target signature data 2 in more random binary trees belonging to destination node be respectively node 2 in random binary tree 6 and with
Node 6 in machine binary tree 8;The target signature data 3 of enterprise 1 destination node affiliated in more random binary trees is random
Node 4 in binary tree 7.
S703, detection target signature data respectively destination node affiliated in every random binary tree whether meet it is default
Risk Production conditions;If satisfied, executing step S604;If not satisfied, executing step S605;
In the embodiment of the present application, detect the target signature data of enterprise respectively in every random binary tree belonging to mesh
The mode whether mark node meets preset risk Production conditions can be with are as follows: determines the target signature data of enterprise respectively at every
Destination node belonging in random binary tree obtains identified depth of the destination node in the random binary tree belonging to it,
The quantity of destination node in destination node determined by obtaining in target predetermined depth, judges whether the quantity is greater than target
Threshold value, if more than, it is determined that target signature data respectively in every random binary tree belonging to destination node meet it is preset
Risk Production conditions, if being not more than, it is determined that target signature data respectively in every random binary tree belonging to destination node
It is unsatisfactory for preset risk Production conditions.
For example, still by taking above-described embodiment as an example, the target signature data 1 of enterprise 1 institute in more random binary trees is determined
The destination node of category be respectively the node 1 in random binary tree 6, the node 3 in random binary tree 7 and random binary tree 8
In node 5;The target signature data 2 of enterprise 1 destination node affiliated in more random binary trees is respectively random y-bend
The node 6 in node 2 and random binary tree 8 in tree 6;The target signature data 3 of enterprise 1 are affiliated in more random binary trees
Destination node be random binary tree 7 in node 4.
If target predetermined depth is depth 2-3, if depth of the node 1 in random binary tree 6 is depth 2, node 2 exists
Depth in random binary tree 6 is depth 3, and depth of the node 3 in random binary tree 7 is depth 2, and node 4 is in random y-bend
Depth in tree 7 is depth 4, and depth of the node 5 in random binary tree 8 is depth 2, depth of the node 6 in random binary tree 8
Degree is depth 3;It is 3 that then the target signature data 1 of enterprise 1, which fall into the quantity of the node of depth 2-3,;The target signature of enterprise 1
The quantity that data 2 fall into the node of depth 2-3 is 2;The target signature data 4 of enterprise 1 fall into the quantity of the node of depth 2-3
It is 0.
If targets threshold is 2, it can determine that the target signature data 1 of enterprise 1 fall into the quantity of the node of depth 2-3
Greater than targets threshold, and then it can determine that dimension belonging to target signature data 1 is that there are the originals of the risk of risk classifications for enterprise 1
Cause;And the target signature data 2 of enterprise 1 fall into the quantity of the node of depth 2-3 no more than targets threshold, then can determine target
Dimension belonging to characteristic 2 is not the reason of there are the risks of risk classifications of enterprise 1;The target signature data 3 of enterprise 1 are fallen into
The quantity of the node of depth 2-3 is not more than targets threshold, then can determine dimension belonging to target signature data 3 not is enterprise 1
The reason of there are the risks of risk classifications.
S704, the reason of dimension belonging to target signature data is determined as risk of the enterprise there are risk classifications;
S705, the reason of determining dimension belonging to target signature data not and be enterprise there are the risks of risk classifications.
A kind of business risk monitoring method provided by the embodiments of the present application, during realizing business risk monitoring not
It needs by the artificial constructed risk monitoring and control rule of expert, therefore can be to avoid existing enterprise's risk monitoring and control technology because needing the artificial structure of expert
Build risk monitoring and control rule, caused by time-consuming and laborious problem.
The embodiment of the present application can also provide a kind of enterprise's wind on the basis of the above-mentioned business risk monitoring method of offer
Dangerous monitoring device is a kind of structural schematic diagram of business risk monitoring device provided by the embodiments of the present application referring to Fig. 8.
As shown in figure 8, the business risk monitoring device includes:
Information determination unit 81, at least one enterprise and at least one enterprise for determining pending monitoring are pending
The risk classifications of monitoring;
Characteristic acquiring unit 82, for obtaining enterprise each of at least one matched dimension of risk classifications
The characteristic of dimension;
Risk monitoring and control unit 83, for combining at least one enterprise every at least one dimension based on isolated forest algorithm
The characteristic of a dimension carries out risk monitoring and control at least one enterprise, obtains risk monitoring and control result.
Further, a kind of risk monitoring and control device provided by the embodiments of the present application further include:
Cut-point determination unit, each dimension for being directed at least one dimension, according to every at least one enterprise
A enterprise determines the cut-point of the dimension in the characteristic of the dimension;
Characteristic amending unit corrects at least one enterprise each enterprise in dimension for the cut-point using dimension
Characteristic;
Correspondingly, risk monitoring and control unit is specifically used for combining at least one enterprise at least one based on isolated forest algorithm
The revised characteristic of each dimension in dimension carries out risk monitoring and control at least one enterprise, obtains risk monitoring and control result.
In the embodiment of the present application, it is preferred that cut-point determination unit includes:
First statistic unit is greater than 0 enterprise in the characteristic of dimension for counting at least one enterprise, obtains the
One number of the enterprise;
Second statistic unit obtains for counting at least one enterprise in enterprise of the characteristic of dimension less than 0
Two numbers of the enterprise;
Judging unit, for judging whether dimension is target dimension;
First cut-point determines subelement, if being target dimension for dimension, is based on the first number of the enterprise and the second enterprise
Quantity calculates the cut-point of dimension;
Second cut-point determines subelement, if not being target dimension for dimension, based on the first number of the enterprise and the second enterprise
Industry quantity corrects the prior distribution of target dimension, obtains the cut-point of dimension.
In the embodiment of the present application, it is preferred that characteristic amending unit includes:
Whether first detection unit is greater than the cut-point of dimension for detecting enterprise in the characteristic of dimension;
Stick unit retains enterprise in dimension if being not more than the cut-point of dimension in the characteristic of dimension for enterprise
Characteristic;
Amending unit, if for enterprise dimension characteristic be greater than dimension cut-point, by enterprise dimension spy
Sign data are set to preset value;
Processing unit, for each enterprise at least one enterprise to be normalized in the characteristic of dimension,
Each enterprise is obtained at least one enterprise in the target signature data of dimension.
In the embodiment of the present application, it is preferred that risk monitoring and control unit includes:
Random binary tree generation unit, for being based on isolating forest algorithm at least one enterprise at least one dimension
The revised characteristic of each dimension is split, and obtains random binary tree;
Leaf node analytical unit, for repeatedly isolating forest in operation for each enterprise at least one enterprise
In algorithm obtained more random binary trees, the leaf node that enterprise falls into every random binary tree is counted;
Risk monitoring and control subelement, the leaf node for being fallen into every random binary tree according to enterprise, determines enterprise
With the presence or absence of the risk of risk classifications.
In the embodiment of the present application, it is preferred that risk monitoring and control subelement, comprising:
Leaf node statistic unit, the leaf node for being fallen into every random binary tree according to enterprise, statistics enterprise
Industry falls into the number of the leaf node of same depth in more random binary trees;
Judging unit, for judging that the number of leaf node that enterprise falls into predetermined depth in more random binary trees is
It is no to meet the preset and matched risk conditions of risk classifications;
First determination unit, for there are the risks of risk classifications if satisfied, determining enterprise;
Second determination unit, for if not satisfied, determining that the risk of risk classifications is not present in enterprise.
Further, a kind of business risk monitoring device provided by the embodiments of the present application further includes reason positioning unit, should
Reason positioning unit includes:
Destination node determination unit, in risk monitoring and control result characterization enterprise there are when the risk of risk classifications, for
Every random binary tree, after obtaining the revised characteristic for the enterprise that random binary tree uses when generating and correcting
Characteristic belonging to destination node, destination node belonging to revised characteristic is in random binary tree based on amendment
The node that the affiliated dimension of characteristic afterwards is split at least one enterprise in the revised characteristic of dimension;
Destination node statistic unit, for after counting and using each amendment of enterprise when more random binary trees generate
Characteristic and each revised characteristic respectively in each random binary tree belonging to destination node;
Second detection unit, for detect revised characteristic respectively in every random binary tree belonging to target
Whether node meets preset risk Production conditions;
Third determination unit, for there are wind if satisfied, dimension belonging to revised characteristic is determined as enterprise
The reason of risk of dangerous type;
4th determination unit, for if not satisfied, determining that dimension belonging to revised characteristic is not that enterprise exists
The reason of risk of risk classifications.
Further, the embodiment of the present application also provides a kind of computer readable storage medium, the computer-readable storage medium
Computer executable instructions are stored in matter, the computer executable instructions are for executing above-mentioned business risk monitoring method.
Optionally, the refinement function of computer executable instructions and extension function can refer to above description.
The application provides a kind of business risk monitoring method, device, server and storage medium, this method by determine to
The risk classifications of at least one enterprise being monitored and the pending monitoring of at least one enterprise, and be directed to and the risk class
Each dimension at least one matched dimension of type obtains at least one enterprise each enterprise in the feature of the dimension respectively
Data, and then each enterprise each dimension at least one dimension at least one enterprise is combined based on isolated forest algorithm
Characteristic carries out the mode of risk monitoring and control at least one enterprise, and available risk monitoring and control result is to realize to business risk
Monitoring.
The application does not need during realizing business risk monitoring by expert's artificial constructed risk monitoring and control rule, because
This can to avoid existing enterprise's risk monitoring and control technology because needing expert's artificial constructed risk monitoring and control rule, caused by it is time-consuming and laborious
Problem.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments in the case where not departing from core of the invention thought or scope.Therefore, originally
Invention is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein
Consistent widest scope.
Claims (10)
1. a kind of business risk monitoring method characterized by comprising
Determine at least one enterprise of pending monitoring and the risk classifications of the pending monitoring of at least one enterprise;
The enterprise is obtained in the characteristic with each of at least one matched dimension of risk classifications dimension;
Based on isolated forest algorithm in conjunction with the spy of at least one described enterprise each dimension at least one described dimension
Data are levied, risk monitoring and control is carried out at least one described enterprise, obtains risk monitoring and control result.
2. the method according to claim 1, wherein when the risk classifications be financial risk type when, the party
Method further include:
For each of at least one dimension dimension, existed according to the enterprise each at least one described enterprise
The characteristic of the dimension determines the cut-point of the dimension;
Characteristic of each enterprise in the dimension at least one enterprise described in cut-point amendment using the dimension
According to;
It is described based on isolated forest algorithm in conjunction at least one described enterprise each dimension at least one described dimension
Characteristic, risk monitoring and control is carried out at least one described enterprise, obtains risk monitoring and control result, comprising: based on isolated forest
Algorithm in conjunction at least one described enterprise each dimension at least one described dimension revised characteristic, it is right
At least one described enterprise carries out risk monitoring and control, obtains risk monitoring and control result.
3. the method according to claim 1, wherein each enterprise in described at least one enterprise according to
Industry determines the cut-point of the dimension in the characteristic of the dimension, comprising:
The enterprise that the characteristic at least one described enterprise in the dimension is greater than 0 is counted, the first number of the enterprise is obtained;
It counts at least one described enterprise in enterprise of the characteristic of the dimension less than 0, obtains the second number of the enterprise;
Judge whether the dimension is target dimension;
If the dimension is target dimension, it is based on first number of the enterprise and the second number of the enterprise, calculates point of the dimension
Cutpoint;
If the dimension is not target dimension, the target dimension is corrected based on first number of the enterprise and the second number of the enterprise
Prior distribution, obtain the cut-point of the dimension.
4. according to the method described in claim 2, it is characterized in that, described described at least using the cut-point amendment of the dimension
Characteristic of each enterprise in the dimension in one enterprise, comprising:
Detect the cut-point whether enterprise is greater than the dimension in the characteristic of the dimension;
If the enterprise is not more than the cut-point of the dimension in the characteristic of the dimension, retain the enterprise in the dimension
The characteristic of degree;
If the enterprise is greater than the cut-point of the dimension in the characteristic of the dimension, by the enterprise in the dimension
Characteristic is set to preset value;
Each enterprise at least one described enterprise is normalized in the characteristic of the dimension, obtains institute
Each enterprise is stated at least one enterprise in the target signature data of the dimension.
5. according to the method described in claim 2, it is characterized in that, it is described based on isolated forest algorithm in conjunction with it is described at least one
The revised characteristic of enterprise's each dimension at least one described dimension carries out at least one described enterprise
Risk monitoring and control obtains risk monitoring and control result, comprising:
Amendment based on isolated forest algorithm at least one described enterprise each dimension at least one described dimension
Characteristic afterwards is split, and obtains random binary tree;
For the enterprise, each of at least one enterprise, running, the repeatedly isolated forest algorithm is obtained more
In random binary tree, the leaf node that the enterprise falls into every random binary tree is counted;
According to the leaf node that the enterprise falls into every random binary tree, determine the enterprise with the presence or absence of described
The risk of risk classifications.
6. according to the method described in claim 5, it is characterized in that, it is described according to the enterprise in every random binary tree
In the leaf node that falls into, determine that the enterprise whether there is the risk of the risk classifications, comprising:
According to the leaf node that the enterprise falls into every random binary tree, count the enterprise described in more with
The number of the leaf node of same depth is fallen into machine binary tree;
It is pre- to judge whether the number for the leaf node that the enterprise falls into predetermined depth in the more random binary trees meets
If with the matched risk conditions of the risk classifications;
If satisfied, determining the enterprise, there are the risks of the risk classifications;
If not satisfied, determining that the risk of the risk classifications is not present in the enterprise.
7. according to the method described in claim 5, it is characterized in that, there are institutes if the risk monitoring and control result characterizes the enterprise
When stating the risk of risk classifications, this method further include:
For random binary tree described in every, after the amendment for obtaining the enterprise that the random binary tree uses when generating
Characteristic and the revised characteristic belonging to destination node, mesh belonging to the revised characteristic
Mark node is to be existed based on the revised affiliated dimension of characteristic at least one described enterprise in the random binary tree
The node that the revised characteristic of the dimension is split;
Count each revised characteristic that the enterprise is used when described more random binary trees generate and often
A revised characteristic respectively in each random binary tree belonging to destination node;
Detect the revised characteristic respectively in every random binary tree belonging to destination node whether meet
Preset risk Production conditions;
If satisfied, dimension belonging to the revised characteristic is determined as the enterprise, there are the wind of the risk classifications
The reason of danger;
If not satisfied, determining that dimension belonging to the revised characteristic is not that there are the risk classifications for the enterprise
The reason of risk.
8. a kind of business risk monitoring device characterized by comprising
Information determination unit, for determine pending monitoring at least one enterprise and the pending prison of at least one enterprise
The risk classifications of control;
Characteristic acquiring unit, for obtain the enterprise with it is every at least one matched dimension of the risk classifications
The characteristic of a dimension;
Risk monitoring and control unit, for being based on isolating forest algorithm in conjunction at least one described enterprise at least one described dimension
The characteristic of each dimension carries out risk monitoring and control at least one described enterprise, obtains risk monitoring and control result.
9. a kind of server characterized by comprising at least one processor and at least one processor;The memory is deposited
Program is contained, the processor calls the program of the memory storage, and described program is any for realizing such as claim 1-7
Business risk monitoring method described in one.
10. a kind of storage medium, which is characterized in that be stored with computer executable instructions, the calculating in the storage medium
Machine executable instruction requires business risk monitoring method described in 1-7 any one for perform claim.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910517033.6A CN110232525B (en) | 2019-06-14 | 2019-06-14 | Enterprise risk acquisition method, device, server and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910517033.6A CN110232525B (en) | 2019-06-14 | 2019-06-14 | Enterprise risk acquisition method, device, server and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110232525A true CN110232525A (en) | 2019-09-13 |
CN110232525B CN110232525B (en) | 2024-04-26 |
Family
ID=67859196
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910517033.6A Active CN110232525B (en) | 2019-06-14 | 2019-06-14 | Enterprise risk acquisition method, device, server and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110232525B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021120845A1 (en) * | 2019-12-19 | 2021-06-24 | 支付宝(杭州)信息技术有限公司 | Homogeneous risk unit feature set generation method, apparatus and device, and medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108877949A (en) * | 2018-06-11 | 2018-11-23 | 吉林大学 | Down's syndrome screening method based on isolated forest algorithm and voting mechanism |
CN109345137A (en) * | 2018-10-22 | 2019-02-15 | 广东精点数据科技股份有限公司 | A kind of rejecting outliers method based on agriculture big data |
CN109345377A (en) * | 2018-09-28 | 2019-02-15 | 北京九章云极科技有限公司 | A kind of generating date system and Real-time Data Processing Method |
CN109492945A (en) * | 2018-12-14 | 2019-03-19 | 深圳壹账通智能科技有限公司 | Business risk identifies monitoring method, device, equipment and storage medium |
CN109714343A (en) * | 2018-12-28 | 2019-05-03 | 北京天融信网络安全技术有限公司 | A kind of judgment method and device of exception of network traffic |
CN109829628A (en) * | 2019-01-07 | 2019-05-31 | 平安科技(深圳)有限公司 | Method for prewarning risk, device and computer equipment based on big data |
-
2019
- 2019-06-14 CN CN201910517033.6A patent/CN110232525B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108877949A (en) * | 2018-06-11 | 2018-11-23 | 吉林大学 | Down's syndrome screening method based on isolated forest algorithm and voting mechanism |
CN109345377A (en) * | 2018-09-28 | 2019-02-15 | 北京九章云极科技有限公司 | A kind of generating date system and Real-time Data Processing Method |
CN109345137A (en) * | 2018-10-22 | 2019-02-15 | 广东精点数据科技股份有限公司 | A kind of rejecting outliers method based on agriculture big data |
CN109492945A (en) * | 2018-12-14 | 2019-03-19 | 深圳壹账通智能科技有限公司 | Business risk identifies monitoring method, device, equipment and storage medium |
CN109714343A (en) * | 2018-12-28 | 2019-05-03 | 北京天融信网络安全技术有限公司 | A kind of judgment method and device of exception of network traffic |
CN109829628A (en) * | 2019-01-07 | 2019-05-31 | 平安科技(深圳)有限公司 | Method for prewarning risk, device and computer equipment based on big data |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021120845A1 (en) * | 2019-12-19 | 2021-06-24 | 支付宝(杭州)信息技术有限公司 | Homogeneous risk unit feature set generation method, apparatus and device, and medium |
Also Published As
Publication number | Publication date |
---|---|
CN110232525B (en) | 2024-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109615524B (en) | Money laundering crime group partner identification method, money laundering crime group partner identification device, computer equipment and storage medium | |
CN111506485B (en) | Feature binning method, device, equipment and computer-readable storage medium | |
CN108833139B (en) | OSSEC alarm data aggregation method based on category attribute division | |
CN110322295B (en) | Relationship strength determination method and system, server and computer readable medium | |
CN109272402A (en) | Modeling method, device, computer equipment and the storage medium of scorecard | |
CN110189165A (en) | Channel abnormal user and abnormal channel recognition methods and device | |
CN110033348A (en) | Storage, checking method and the device of business datum | |
CN108268477A (en) | A kind of customer relationship based on ticket determines method and apparatus | |
CN112084343A (en) | Method, device and medium for quantifying social relationship graph | |
CN110232525A (en) | A kind of business risk monitoring method, device, server and storage medium | |
CN107330709B (en) | Method and device for determining target object | |
CN109428760B (en) | User credit evaluation method based on operator data | |
CN113505990A (en) | Enterprise risk assessment method and device, electronic equipment and storage medium | |
CN112419074A (en) | Vehicle insurance fraud group identification method and device | |
CN112217908A (en) | Information pushing method and device based on transfer learning and computer equipment | |
JP7081695B2 (en) | Priority determination device, priority determination method, and control program | |
CN113240259B (en) | Rule policy group generation method and system and electronic equipment | |
CN110062096A (en) | A kind of method and device for screening offending user | |
CN111352924B (en) | Method and device for solving data inclination problem | |
CN110300011A (en) | A kind of alarm root is because of localization method, device and computer readable storage medium | |
CN112307307B (en) | Insurance product recommendation method and apparatus | |
CN110458707B (en) | Behavior evaluation method and device based on classification model and terminal equipment | |
CN113989005A (en) | Tax risk enterprise mining method and device | |
CN108804640B (en) | Data grouping method, device, storage medium and equipment based on maximized IV | |
CN110597572B (en) | Service call relation analysis method and computer system |
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 |