CN108876589A - Methods of risk assessment, device, equipment and storage medium - Google Patents
Methods of risk assessment, device, equipment and storage medium Download PDFInfo
- Publication number
- CN108876589A CN108876589A CN201810414291.7A CN201810414291A CN108876589A CN 108876589 A CN108876589 A CN 108876589A CN 201810414291 A CN201810414291 A CN 201810414291A CN 108876589 A CN108876589 A CN 108876589A
- Authority
- CN
- China
- Prior art keywords
- drosophila
- neighbour
- sample data
- fuzzy
- group
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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]
Abstract
The invention discloses a kind of methods of risk assessment, including：Sample data is obtained, and the sample data is normalized；Processing is iterated to the sample data after normalized using drosophila algorithm, obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor；The risk evaluation model based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient；Risk assessment is carried out to data to be assessed based on the risk evaluation model.In the embodiment of the present invention, drosophila algorithm is dissolved into fuzzy k nearest neighbor to the neighbour's number and vague intensity coefficient for determining fuzzy k nearest neighbor, it is few that parameter is arranged in algorithm, the optimal value of neighbour's number and vague intensity coefficient can quickly and accurately be found, to construct the higher risk evaluation model based on fuzzy k nearest neighbor of risk assessment accuracy rate, to improve the accuracy of risk assessment.The present invention also provides a kind of risk assessment device, equipment and storage mediums.
Description
Technical field
The present invention relates to information technology field more particularly to a kind of methods of risk assessment, assessment device, assessment equipment and deposit
Storage media.
Background technique
Under without the situation that pure fiduciary loan constantly heats up is mortgaged, risk assessment becomes in commercial bank risk management very
One of important work plays great influence to the survival and development of business bank.
With the development of information technology, artificial intelligence technology is more and more applied in risk assessment to construct and be based on
The risk evaluation model of artificial intelligence technology.Wherein, neural network has the characteristics that self-organizing, adaptive, self study as one kind
Nonparametric model, it is not only not stringent to the Spreading requirements of sample data, but also have non-linear mapping capability, generalization ability
With higher precision of prediction, thus, be widely used in risk evaluation model building in, such as mode neural network, probabilistic neural
Network, model-naive Bayesian and multi-layer perception (MLP) etc. are widely used in the building of risk evaluation model.
Compared with the risk evaluation model based on statistical analysis, risk evaluation model neural network based is more intuitive, easy
Understand, this nonlinear model classification problem of risk assessment can be better solved.But this risk neural network based is commented
Estimate model not only and have that structure is complicated, there is the problems such as black box property, decision process lack transparency, and network weight parameter
The problems such as determining, training effectiveness is low is difficult to structure.
To solve the above problems, k nearest neighbor and fuzzy k nearest neighbor are widely used in risk assessment, in the wind based on k nearest neighbor
In the assessment of danger, the distance between test sample and neighbour's sample are not accounted for, and is to confer to neighbour's sample with phase
Same weight, so that the assessment accuracy of risk assessment is lower；In the existing risk assessment based on fuzzy k nearest neighbor,
It can not determine neighbour's number and vague intensity coefficient, and cause the assessment accuracy rate of risk assessment lower.
To sum up, the assessment accuracy for how improving risk assessment becomes those skilled in the art's urgent problem to be solved.
Summary of the invention
The embodiment of the invention provides a kind of methods of risk assessment, device, equipment and storage mediums, can quickly, accurately
Ground finds the optimal value of neighbour's number and vague intensity coefficient in fuzzy k nearest neighbor, so that it is higher to construct risk assessment accuracy rate
The risk evaluation model based on fuzzy k nearest neighbor, to improve the accuracy of risk assessment.
The embodiment of the present invention in a first aspect, provide a kind of methods of risk assessment, including：
Sample data is obtained, and the sample data is normalized；
Processing is iterated to the sample data after normalized using drosophila algorithm, obtains the close of fuzzy k nearest neighbor
Adjacent number and vague intensity coefficient；
The risk assessment mould based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient
Type；
Risk assessment is carried out to data to be assessed based on the risk evaluation model.
Further, described that processing is iterated to the sample data after normalized using drosophila algorithm, it obtains
The neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor, including：
Drosophila population size, target step value and the maximum number of iterations of drosophila algorithm are set；
According to the group position of the sample data initialization drosophila group after normalized；
The heading of drosophila individual in the drosophila group is determined according to the group position and the target step value
And flying distance；
Calculate the flavor concentration value based on the heading and postflight each drosophila individual of the flying distance；
Position where flavor concentration is worth maximum drosophila individual is redefined as the group position of drosophila group, and will
The current iteration number of drosophila group increases a several unit；
Whether the current iteration number after judging several units of increase is equal to the maximum number of iterations；
If increasing the current iteration number after a several units is equal to the maximum number of iterations, by current drosophila group
Group position (p, q) in p be determined as neighbour's number, and q is determined as the vague intensity coefficient；
If increasing the current iteration number after a severals units not equal to the maximum number of iterations, return execution according to
The group position and the target step value determine the heading of drosophila individual and flying distance in the drosophila group
Step and subsequent step.
Optionally, the group position according to the sample data initialization drosophila group after normalized, including：
The group position of formula initialization drosophila group is determined according to following positions：
Wherein, X0For the abscissa of group position, Y0For the ordinate of group position, rand is the random life in [0,1] section
At number, XmaxFor the maximum value of abscissa in sample data after normalized, XminIt is horizontal in sample data after normalized
The minimum value of coordinate, YmaxFor the maximum value of ordinate in sample data after normalized, YminFor sample after normalized
The minimum value of ordinate in data.
Preferably, described that drosophila individual in the drosophila group is determined according to the group position and the target step value
Heading and flying distance, including：
The heading and flying distance of drosophila individual are determined according to following flight equations：
Wherein, (Xi, Yi) it is position where after the flight of i-th drosophila individual, i ∈ [1, R], R are drosophila population size,
Step is target step value, and rand is the random generation number in [0,1] section；
Correspondingly, described to calculate based on the heading and the flying distance postflight each drosophila individual
Flavor concentration value, including：
The flavor concentration value of each drosophila individual is calculated according to following flavor concentration value calculation formula：
Wherein, SiFor the flavor concentration value of i-th of drosophila individual, DiDistance for i-th of drosophila individual to origin, (Xi,
Yi) it is position where after the flight of i-th drosophila individual.
Further, described that the fuzzy k nearest neighbor is based on according to neighbour's number and vague intensity coefficient building
Risk evaluation model, including：
The fuzzy membership of the sample data is determined according to neighbour's number；
It is close based on the fuzzy K according to the building of neighbour's number, the vague intensity coefficient and the fuzzy membership
Adjacent risk evaluation model.
Optionally, the fuzzy membership that the sample data is determined according to neighbour's number, including：
The fuzzy membership of the sample data is determined using following fuzzy membership calculation formula：
Wherein, uij(xj) it is sample data xjIt is under the jurisdiction of the fuzzy membership of the i-th class, njFor sample data xjIt is under the jurisdiction of i-th
Neighbour's sample number of class, k are neighbour's number.
Preferably, described to be based on according to the building of neighbour's number, the vague intensity coefficient and the fuzzy membership
The risk evaluation model of the fuzzy k nearest neighbor, including：
The risk evaluation model based on the fuzzy k nearest neighbor is constructed using following model construction formula:
Wherein, C (x) is the generic of data x to be assessed, and C is classification sum, and k is neighbour's number, and m is vague intensity
Coefficient, | | x-xj| | it is the sample data x of data x to be assessed and its neighbourjBetween Euclidean distance, uij(xj) it is sample data
xjIt is under the jurisdiction of the fuzzy membership of the i-th class.
The second aspect of the embodiment of the present invention provides a kind of risk assessment device, including：
Data normalization module is normalized for obtaining sample data, and to the sample data；
Drosophila algorithm processing module, for being iterated place to the sample data after normalized using drosophila algorithm
Reason obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor；
Assessment models construct module, for being based on the mould according to neighbour's number and vague intensity coefficient building
Paste the risk evaluation model of k nearest neighbor；
Risk evaluation module, for carrying out risk assessment to data to be assessed based on the risk evaluation model.
The third aspect of the embodiment of the present invention, provides a kind of risk assessment equipment, including memory, processor and deposits
The computer program that can be run in the memory and on the processor is stored up, the processor executes the computer journey
It is realized when sequence as described in aforementioned first aspect the step of methods of risk assessment.
The fourth aspect of the embodiment of the present invention, provides a kind of computer readable storage medium, described computer-readable to deposit
Storage media is stored with computer program, and the risk as described in aforementioned first aspect is realized when the computer program is executed by processor
The step of appraisal procedure.
As can be seen from the above technical solutions, the embodiment of the present invention has the following advantages that：
In the embodiment of the present invention, firstly, obtaining sample data, and the sample data is normalized；Secondly,
Processing is iterated to the sample data after normalized using drosophila algorithm, obtain fuzzy k nearest neighbor neighbour's number and
Vague intensity coefficient；Then, the wind based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient
Dangerous assessment models, and risk assessment is carried out to data to be assessed based on the risk evaluation model.In the embodiment of the present invention, by fruit
Fly algorithm is dissolved into fuzzy k nearest neighbor the neighbour's number and vague intensity coefficient for determining fuzzy k nearest neighbor, and parameter is arranged in algorithm
It is few, the optimal value of neighbour's number and vague intensity coefficient can be quickly and accurately found, to construct risk assessment accuracy rate more
The high risk evaluation model based on fuzzy k nearest neighbor, to improve the accuracy of risk assessment.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only of the invention some
Embodiment for those of ordinary skill in the art without any creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is a kind of method flow diagram for methods of risk assessment that the embodiment of the present invention one provides；
Fig. 2 obtains fuzzy k nearest neighbor for methods of risk assessment a kind of in the embodiment of the present invention one under an application scenarios
The flow diagram of neighbour's number and vague intensity coefficient；
Fig. 3 is a kind of structural schematic diagram of risk assessment device provided by Embodiment 2 of the present invention；
Fig. 4 is a kind of structural schematic diagram for risk assessment equipment that the embodiment of the present invention three provides.
Specific embodiment
The embodiment of the invention provides a kind of methods of risk assessment, device, equipment and storage mediums, for quickly, accurately
Ground determines the optimal value of neighbour's number and vague intensity coefficient in fuzzy k nearest neighbor, higher to construct risk assessment accuracy rate
Based on the risk evaluation model of fuzzy k nearest neighbor, to improve the accuracy of risk assessment.
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with the present invention
Attached drawing in embodiment, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that disclosed below
Embodiment be only a part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field
Those of ordinary skill's all other embodiment obtained without making creative work, belongs to protection of the present invention
Range.
Fuzzy k nearest neighbor in the embodiment of the present invention is improved by k nearest neighbor, using the concept of fuzzy set, to difference
Classification distributes degree of membership, meanwhile, the distance of k neighbour is considered, is assigned with different persons in servitude for the sample of each classification
Category degree, so that from the closer sample of data to be assessed than there is bigger degree of membership from the farther away sample of data to be assessed, finally
Classification with highest degree of membership can be used as the final classification of data to be assessed.
Referring to Fig. 1, the embodiment of the present invention one provides a kind of methods of risk assessment, the methods of risk assessment, including：
Step S101, sample data is obtained, and the sample data is normalized.
In the present embodiment, a certain number of sample datas, such as enterprise operation data, personal credit's data are obtained first,
And known to the attributive character and class label of acquired sample data.After obtaining these sample datas, to each
The attributive character of sample data is normalized, and characteristic value corresponding to each sample data is mapped to [0,1]
It in section, is disturbed caused by smaller characteristic value to avoid larger characteristic value, reduces big data sample value to small data sample value
It must disturb, the sample data after normalized is enabled effectively to support the treatment process of drosophila algorithm.
Specifically, in the present embodiment, normalizing is carried out according to attributive character of following normalization formula to the sample data
Change processing:
Wherein, tc(n) ' is the characteristic value after n-th of attributive character normalized of each sample data, tcIt (n) is every
The initial characteristic values of n-th of attributive character of one sample data, t (n)maxFor the maximum value of n-th of attributive character, t (n)minFor
The minimum value of n-th of attributive character, n ∈ [1, N], N are the attributive character that sample data includes, and c ∈ [1, C], C are that classification is total
Number.
Step S102, processing is iterated to the sample data after normalized using drosophila algorithm, obtains fuzzy K
The neighbour's number and vague intensity coefficient of neighbour.
In the present embodiment, after sample data is normalized, using drosophila algorithm to after normalized
Sample data be iterated processing, to determine the neighbour's number and vague intensity of fuzzy k nearest neighbor according to iterative processing result
Coefficient, thus neighbour's number and vague intensity coefficient after being optimized, wherein neighbour's number after optimization is closed on for determination
The quantity of sample, distance of the vague intensity coefficient based on neighbour after optimization are that the sample data of each classification is respectively different
Degree of membership, the sample data close from data to be assessed have bigger degree of membership than the sample data remote from data to be assessed.
Here, for ease of understanding, according to Fig. 1 described embodiment, below with a practical application scene to the present invention
One of embodiment one methods of risk assessment obtains the neighbour's number and vague intensity of fuzzy k nearest neighbor under an application scenarios
Coefficient is described：
As shown in Fig. 2, under this scene, it is described that the sample data after normalized is iterated using drosophila algorithm
Processing obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor, including：
Step S201, drosophila population size, target step value and the maximum number of iterations of drosophila algorithm are set；
Step S202, according to the group position of the sample data initialization drosophila group after normalized；
Step S203, drosophila individual in the drosophila group is determined according to the group position and the target step value
Heading and flying distance；
Step S204, the taste based on the heading and postflight each drosophila individual of the flying distance is calculated
Road concentration value；
Step S205, the position where flavor concentration being worth maximum drosophila individual is redefined as the group of drosophila group
Position, and the current iteration number of drosophila group is increased into a several unit；
Step S206, whether the current iteration number after judging several units of increase is equal to the maximum number of iterations；
If step S207, increasing the current iteration number after a several units is equal to the maximum number of iterations, will work as
P in the group position (p, q) of preceding drosophila group is determined as neighbour's number, and q is determined as the vague intensity coefficient；
If increasing the current iteration number after a several units is not equal to the maximum number of iterations, execution is returned according to the group
The step of position and the target step value determine the heading and flying distance of drosophila individual in the drosophila group and
Subsequent step.
For above-mentioned steps S201, the iterative processing that drosophila algorithm is carried out to the sample data after normalized it
Before, need to carry out Initialize installation to the parameter that is related in drosophila algorithm, as be arranged drosophila population size R in drosophila algorithm,
Target step value Step and maximum number of iterations T, wherein target target step value Step is the maximum that drosophila individual deviates every time
Distance.
For above-mentioned steps S202, it is to be understood that the iterative processing of drosophila algorithm removes it needs to be determined that drosophila group advises
Outside mould R, target step value Step and maximum number of iterations T, it is also necessary to the group position of drosophila group is initialized, thus, this
Jing Zhong initializes the group position of drosophila group according to the sample data after normalized, and all drosophila individuals are with this
Group position is that starting point is flown everywhere, to search out food position.In this scene, specifically, determined according to following positions
The group position of formula initialization drosophila group：
Wherein, X0For the abscissa of group position, Y0For the ordinate of group position, rand is the random life in [0,1] section
At number, XmaxFor the maximum value of abscissa in sample data after normalized, XminIt is horizontal in sample data after normalized
The minimum value of coordinate, YmaxFor the maximum value of ordinate in sample data after normalized, YminFor sample after normalized
The minimum value of ordinate in data.
For above-mentioned steps S203, after the group position that drosophila group has been determined, using the group position as starting point, controllably
Each drosophila individual is made with specific heading and flying distance to being flown everywhere with search of food position.It, can in this scene
The heading of drosophila individual and flight in the drosophila group are determined according to the group position and the target step value
Distance specifically determines the heading and flying distance of drosophila individual according to following flight equations：
Wherein, (Xi, Yi) it is position where after the flight of i-th drosophila individual, i ∈ [1, R], R are drosophila population size,
Step is target step value, and rand is the random generation number in [0,1] section.
For above-mentioned steps S204, fly everywhere controlling each drosophila individual according to specific heading and flying distance
After row, it can be learnt that the position where after the flight of each drosophila individual, so as to according to place after the flight of each drosophila individual
Position calculates the flavor concentration value of each drosophila individual, wherein the flavor concentration value of drosophila individual is where drosophila individual
Position to origin distance inverse, i.e., in this scene, calculate each specifically according to following flavor concentration value calculation formula
The flavor concentration value of drosophila individual：
Wherein, SiFor the flavor concentration value of i-th of drosophila individual, DiDistance for i-th of drosophila individual to origin, (Xi,
Yi) it is position where after the flight of i-th drosophila individual.
For above-mentioned steps S205, in this scene, using the flavor concentration value of drosophila individual as the judgement letter of drosophila algorithm
Number.After the flavor concentration value for obtaining each drosophila individual, this multiple flavor concentration value is compared, such as according to multiple tastes
The size of concentration value carries out sequence from large to small.When maximum flavor concentration value is greater than the current flavor concentration value of drosophila group
When, then maximum flavor concentration value is determined as to the current flavor concentration value of drosophila group, and maximum flavor concentration value is corresponding
Position where drosophila individual is determined as the group position of drosophila group, while the current iteration number of drosophila group is increased by one
Number unit.
It is understood that in this scene, when initial, the current flavor concentration value of drosophila group can according to it is initial when drosophila
The group position of group is calculated, i.e., the group position of drosophila group when the initial current flavor concentration value of drosophila group is initial
Set the inverse at a distance from origin.In addition, the current iteration number of drosophila group may be configured as 0 time when initial, and when according to each
It, can be correspondingly by the current iteration of drosophila group when the flight course of a drosophila individual obtains the group position of new drosophila group
Number increases a several unit, such as increases primary.
For above-mentioned steps S206, after the current iteration number of drosophila group is increased a several units, then further
Whether the current iteration number after judging several units of increase is equal to the maximum number of iterations, it is assumed that the greatest iteration time
Number is 50 times, that is, judges whether current iteration number reaches 50 times, if reaching 50 times, then at the iteration that stops drosophila algorithm
Reason；If not up to 50 times, then the iterative processing for continuing to execute drosophila algorithm is needed.
For above-mentioned steps S207, if judging, the current iteration number of drosophila group is equal to the maximum number of iterations
Words, it is determined that the iterative process of drosophila algorithm terminates, and the position by the drosophila with maximum flavor concentration value where individual
The group position for being determined as current drosophila group is set, the optimum population position (p, q) of drosophila group is obtained with this, meanwhile, it will most
P value in excellent group position (p, q) is determined as neighbour's number of fuzzy k nearest neighbor, and q value is determined as the fuzzy of fuzzy k nearest neighbor
Strength factor, with the neighbour's number k and vague intensity Coefficient m of this Optimization of Fuzzy k nearest neighbor；If judging, the current of drosophila group changes
Generation number is not equal to the maximum number of iterations, then the position where maximum flavor concentration being worth corresponding drosophila individual is determined as
Continue to be iterated processing after the group position of drosophila group, it is until reaching maximum number of iterations, i.e., maximum taste is dense
Position where the corresponding drosophila individual of angle value returns to execution according to the group position after being determined as the group position of drosophila group
Set the step of determining the heading and flying distance of drosophila individual in the drosophila group with the target step value and after
Continuous step.
Step S103, the wind based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient
Dangerous assessment models.
After obtaining neighbour's number k and the vague intensity Coefficient m after fuzzy k nearest neighbor optimization, it can construct based on mould
Paste the risk evaluation model of k nearest neighbor.Further, in this embodiment described according to neighbour's number and the vague intensity
Coefficient constructs the risk evaluation model based on the fuzzy k nearest neighbor, specifically includes：Step a, institute is determined according to neighbour's number
State the fuzzy membership of sample data；Step b, according to neighbour's number, the vague intensity coefficient and the fuzzy membership
Risk evaluation model of the degree building based on the fuzzy k nearest neighbor.
For above-mentioned steps a, the fuzzy membership of the sample data is determined according to neighbour's number, is specifically included：
The fuzzy membership of the sample data is determined using following fuzzy membership calculation formula：
Wherein, uij(xj) it is sample data xjIt is under the jurisdiction of the fuzzy membership of the i-th class, njFor sample data xjIt is under the jurisdiction of i-th
Neighbour's sample number of class, k are neighbour's number.
For above-mentioned steps b, base is constructed according to neighbour's number, the vague intensity coefficient and the fuzzy membership
In the risk evaluation model of the fuzzy k nearest neighbor, specifically include：
The risk evaluation model based on the fuzzy k nearest neighbor is constructed using following model construction formula:
Wherein, C (x) is the generic of data x to be assessed, and C is classification sum, and k is neighbour's number, and m is vague intensity
Coefficient, | | x-xj| | it is the sample data x of data x to be assessed and its neighbourjBetween Euclidean distance, uij(xj) it is sample data
xjIt is under the jurisdiction of the fuzzy membership of the i-th class.
Step S104, risk assessment is carried out to data to be assessed based on the risk evaluation model.
In the present embodiment, after obtaining the risk evaluation model based on fuzzy k nearest neighbor, risk is carried out to data to be assessed
When assessment, it is input in the risk evaluation model, can be obtained described to be evaluated using the data to be assessed as input data
Estimate the risk evaluation result of data.
In the embodiment of the present invention, firstly, obtaining sample data, and the sample data is normalized；Secondly,
Processing is iterated to the sample data after normalized using drosophila algorithm, obtain fuzzy k nearest neighbor neighbour's number and
Vague intensity coefficient；Then, the wind based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient
Dangerous assessment models, and risk assessment is carried out to data to be assessed based on the risk evaluation model.In the embodiment of the present invention, by fruit
Fly algorithm is dissolved into fuzzy k nearest neighbor the neighbour's number and vague intensity coefficient for determining fuzzy k nearest neighbor, and parameter is arranged in algorithm
It is few, the optimal value of neighbour's number and vague intensity coefficient can be quickly and accurately found, to construct risk assessment accuracy rate more
The high risk evaluation model based on fuzzy k nearest neighbor, to improve the accuracy of risk assessment.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
A kind of methods of risk assessment is essentially described above, a kind of risk assessment device will be described in detail below.
Fig. 3 shows a kind of structural schematic diagram of risk assessment device provided by Embodiment 2 of the present invention.As shown in figure 3,
The risk assessment device, including：
Data normalization module 301 is normalized for obtaining sample data, and to the sample data；
Drosophila algorithm processing module 302, for being changed using drosophila algorithm to the sample data after normalized
Generation processing obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor；
Assessment models construct module 303, for being based on institute according to neighbour's number and vague intensity coefficient building
State the risk evaluation model of fuzzy k nearest neighbor；
Risk evaluation module 304, for carrying out risk assessment to data to be assessed based on the risk evaluation model.
Further, the drosophila algorithm processing module 302, including：
Setting unit, for drosophila population size, target step value and the maximum number of iterations of drosophila algorithm to be arranged；
Initialization unit, for the group position according to the sample data initialization drosophila group after normalized；
Flight determination unit, for determining fruit in the drosophila group according to the group position and the target step value
The heading and flying distance of fly individual；
Flavor concentration value computing unit, it is postflight each based on the heading and the flying distance for calculating
The flavor concentration value of a drosophila individual；
The number of iterations adding unit is redefined as fruit for the position where flavor concentration is worth maximum drosophila individual
The group position of fly group, and the current iteration number of drosophila group is increased into a several unit；
The number of iterations judging unit, for judge increase a severals units after current iteration number whether be equal to described in most
Big the number of iterations；
Factor determination unit, if being equal to the greatest iteration time for increasing the current iteration number after a several units
P in the group position (p, q) of current drosophila group is then determined as neighbour's number, and q is determined as described obscure by number
Strength factor；
Execution unit is returned to, if to be not equal to the greatest iteration secondary for increasing the current iteration number after severals units
Number then returns to the flight for executing and determining drosophila individual in the drosophila group according to the group position and the target step value
The step of direction and flying distance and subsequent step.
Optionally, the initialization unit, specifically for determining the group of formula initialization drosophila group according to following positions
Body position：
Wherein, X0For the abscissa of group position, Y0For the ordinate of group position, rand is the random life in [0,1] section
At number, XmaxFor the maximum value of abscissa in sample data after normalized, XminIt is horizontal in sample data after normalized
The minimum value of coordinate, YmaxFor the maximum value of ordinate in sample data after normalized, YminFor sample after normalized
The minimum value of ordinate in data.
Preferably, the flight determination unit, specifically for determining the flight side of drosophila individual according to following flight equations
To and flying distance：
Wherein, (Xi, Yi) it is position where after the flight of i-th drosophila individual, i ∈ [1, R], R are drosophila population size,
Step is target step value, and rand is the random generation number in [0,1] section；
Correspondingly, the flavor concentration value computing unit is specifically used for being calculated according to following flavor concentration value calculation formula
The flavor concentration value of each drosophila individual：
Wherein, SiFor the flavor concentration value of i-th of drosophila individual, DiDistance for i-th of drosophila individual to origin, (Xi,
Yi) it is position where after the flight of i-th drosophila individual.
Further, the assessment models construct module 303, including：
Fuzzy membership determination unit, for determining the fuzzy membership of the sample data according to neighbour's number；
Assessment models construction unit, for according to neighbour's number, the vague intensity coefficient and the fuzzy membership
Risk evaluation model of the degree building based on the fuzzy k nearest neighbor.
Optionally, the fuzzy membership determination unit is specifically used for determining using following fuzzy membership calculation formula
The fuzzy membership of the sample data：
Wherein, uij(xj) it is sample data xjIt is under the jurisdiction of the fuzzy membership of the i-th class, njFor sample data xjIt is under the jurisdiction of i-th
Neighbour's sample number of class, k are neighbour's number.
Preferably, the assessment models construction unit is specifically used for using the building of following model construction formula based on described
The risk evaluation model of fuzzy k nearest neighbor:
Wherein, C (x) is the generic of data x to be assessed, and C is classification sum, and k is neighbour's number, and m is vague intensity
Coefficient, | | x-xj| | it is the sample data x of data x to be assessed and its neighbourjBetween Euclidean distance, uij(xj) it is sample data
xjIt is under the jurisdiction of the fuzzy membership of the i-th class.
Fig. 4 is the schematic diagram for the risk assessment equipment that the embodiment of the present invention three provides.As shown in figure 4, the wind of the embodiment
Dangerous assessment equipment 400 includes：It processor 401, memory 402 and is stored in the memory 402 and can be in the processing
The computer program 403 run on device 401, such as risk assessment procedures.The processor 401 executes the computer program
The step in above-mentioned each methods of risk assessment embodiment, such as step S101 shown in FIG. 1 to step S104 are realized when 403.
Alternatively, the processor 401 realizes each module/unit in above-mentioned each Installation practice when executing the computer program 403
Function, such as module shown in Fig. 3 301 is to the function of module 304.
Illustratively, the computer program 403 can be divided into one or more module/units, it is one or
Multiple module/the units of person are stored in the memory 402, and are executed by the processor 401, to complete the present invention.Institute
Stating one or more module/units can be the series of computation machine program instruction section that can complete specific function, the instruction segment
For describing implementation procedure of the computer program 403 in the risk assessment equipment 400.For example, the computer journey
Sequence 403 can be divided into data normalization module, drosophila algorithm processing module, assessment models building module, risk assessment mould
Block, each module concrete function are as follows：
Data normalization module is normalized for obtaining sample data, and to the sample data；
Drosophila algorithm processing module, for being iterated place to the sample data after normalized using drosophila algorithm
Reason obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor；
Assessment models construct module, for being based on the mould according to neighbour's number and vague intensity coefficient building
Paste the risk evaluation model of k nearest neighbor；
Risk evaluation module, for carrying out risk assessment to data to be assessed based on the risk evaluation model.
The risk assessment equipment 400 can be the meter such as desktop PC, notebook, palm PC and cloud server
Calculate equipment.The risk assessment equipment 400 may include, but be not limited only to, processor 401, memory 402.Those skilled in the art
Member is appreciated that Fig. 4 is only the example of risk assessment equipment 400, does not constitute the restriction to risk assessment equipment 400, can
To include perhaps combining certain components or different components, such as the risk assessment than illustrating more or fewer components
Equipment 400 can also include input-output equipment, network access equipment, bus etc..
The processor 401 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng.
The memory 402 can be the internal storage unit of the risk assessment equipment 400, such as risk assessment equipment
400 hard disk or memory.The memory 402 is also possible to the External memory equipment of the risk assessment equipment 400, such as institute
State the plug-in type hard disk being equipped in risk assessment equipment 400, intelligent memory card (Smart Media Card, SMC), secure digital
(Secure Digital, SD) card, flash card (Flash Card) etc..Further, the memory 402 can also both include
The internal storage unit of the risk assessment equipment 400 also includes External memory equipment.The memory 402 is described for storing
Other programs and data needed for computer program and the risk assessment equipment 400.The memory 402 can be also used for
Temporarily store the data that has exported or will export.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, is not described in detail or remembers in some embodiment
The part of load may refer to the associated description of other embodiments.
Those of ordinary skill in the art may be aware that each embodiment described in conjunction with the examples disclosed in this document
Module, unit and/or method and step can be realized with the combination of electronic hardware or computer software and electronic hardware.This
A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not
It is considered as beyond the scope of this invention.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, the indirect coupling of device or unit
It closes or communicates to connect, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the present invention realizes above-described embodiment side
All or part of the process in method can also instruct relevant hardware to complete, the computer by computer program
Program can be stored in a computer readable storage medium, and the computer program is when being executed by processor, it can be achieved that above-mentioned each
The step of a embodiment of the method.Wherein, the computer program includes computer program code, and the computer program code can
Think source code form, object identification code form, executable file or certain intermediate forms etc..The computer-readable medium can be with
Including：Any entity or device, recording medium, USB flash disk, mobile hard disk, magnetic disk, light of the computer program code can be carried
Disk, computer storage, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random
Access Memory), electric carrier signal, telecommunication signal and software distribution medium etc..It should be noted that the computer
The content that readable medium includes can carry out increase and decrease appropriate according to the requirement made laws in jurisdiction with patent practice, such as
It does not include electric carrier signal and telecommunication signal according to legislation and patent practice, computer-readable medium in certain jurisdictions.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although referring to before
Stating embodiment, invention is explained in detail, those skilled in the art should understand that：It still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of methods of risk assessment, which is characterized in that including：
Sample data is obtained, and the sample data is normalized；
Processing is iterated to the sample data after normalized using drosophila algorithm, obtains the neighbour of fuzzy k nearest neighbor
Several and vague intensity coefficient；
The risk evaluation model based on the fuzzy k nearest neighbor is constructed according to neighbour's number and the vague intensity coefficient；
Risk assessment is carried out to data to be assessed based on the risk evaluation model.
2. methods of risk assessment according to claim 1, which is characterized in that it is described using drosophila algorithm at through normalization
Sample data after reason is iterated processing, obtains the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor, including：
Drosophila population size, target step value and the maximum number of iterations of drosophila algorithm are set；
According to the group position of the sample data initialization drosophila group after normalized；
The heading of drosophila individual in the drosophila group is determined according to the group position and the target step value and is flown
Row distance；
Calculate the flavor concentration value based on the heading and postflight each drosophila individual of the flying distance；
Position where flavor concentration is worth maximum drosophila individual is redefined as the group position of drosophila group, and by drosophila
The current iteration number of group increases a several unit；
Whether the current iteration number after judging several units of increase is equal to the maximum number of iterations；
If increasing the current iteration number after a several units is equal to the maximum number of iterations, by the group of current drosophila group
P in body position (p, q) is determined as neighbour's number, and q is determined as the vague intensity coefficient；
If increasing the current iteration number after a several units is not equal to the maximum number of iterations, execution is returned according to
The step of group position and the target step value determine the heading and flying distance of drosophila individual in the drosophila group
And subsequent step.
3. methods of risk assessment according to claim 2, which is characterized in that the sample number according to after normalized
According to the group position of initialization drosophila group, including：
The group position of formula initialization drosophila group is determined according to following positions：
Wherein, X0For the abscissa of group position, Y0For the ordinate of group position, rand is the random generation in [0,1] section
Number, XmaxFor the maximum value of abscissa in sample data after normalized, XminFor horizontal seat in sample data after normalized
Target minimum value, YmaxFor the maximum value of ordinate in sample data after normalized, YminFor sample number after normalized
According to the minimum value of middle ordinate.
4. methods of risk assessment according to claim 3, which is characterized in that described according to the group position and the mesh
Mark step value determines the heading and flying distance of drosophila individual in the drosophila group, including：
The heading and flying distance of drosophila individual are determined according to following flight equations：
Wherein, (Xi, Yi) it is position where after the flight of i-th drosophila individual, i ∈ [1, R], R are drosophila population size, Step
For target step value, rand is the random generation number in [0,1] section；
Correspondingly, the taste of the calculating based on the heading and postflight each drosophila individual of the flying distance
Concentration value, including：
The flavor concentration value of each drosophila individual is calculated according to following flavor concentration value calculation formula：
Wherein, SiFor the flavor concentration value of i-th of drosophila individual, DiDistance for i-th of drosophila individual to origin, (Xi, Yi) be
Position where after i-th of drosophila individual flight.
5. methods of risk assessment according to any one of claim 1 to 4, which is characterized in that described according to the neighbour
Number and the vague intensity coefficient construct the risk evaluation model based on the fuzzy k nearest neighbor, including：
The fuzzy membership of the sample data is determined according to neighbour's number；
It is constructed according to neighbour's number, the vague intensity coefficient and the fuzzy membership based on the fuzzy k nearest neighbor
Risk evaluation model.
6. methods of risk assessment according to claim 5, which is characterized in that described according to neighbour's number determination
The fuzzy membership of sample data, including：
The fuzzy membership of the sample data is determined using following fuzzy membership calculation formula：
Wherein, uij(xj) it is sample data xjIt is under the jurisdiction of the fuzzy membership of the i-th class, njFor sample data xjIt is under the jurisdiction of the i-th class
Neighbour's sample number, k are neighbour's number.
7. methods of risk assessment according to claim 6, which is characterized in that described according to neighbour's number, the mould
It pastes strength factor and the fuzzy membership constructs the risk evaluation model based on the fuzzy k nearest neighbor, including：
The risk evaluation model based on the fuzzy k nearest neighbor is constructed using following model construction formula：
Wherein, C (x) is the generic of data x to be assessed, and C is classification sum, and k is neighbour's number, and m is vague intensity coefficient,
||x-xj| | it is the sample data x of data x to be assessed and its neighbourjBetween Euclidean distance, uij(xj) it is sample data xjIt is subordinate to
In the fuzzy membership of the i-th class.
8. a kind of risk assessment device, which is characterized in that including：
Data normalization module is normalized for obtaining sample data, and to the sample data；
Drosophila algorithm processing module, for being iterated processing to the sample data after normalized using drosophila algorithm,
Obtain the neighbour's number and vague intensity coefficient of fuzzy k nearest neighbor；
Assessment models construct module, for being based on the fuzzy K according to neighbour's number and vague intensity coefficient building
The risk evaluation model of neighbour；
Risk evaluation module, for carrying out risk assessment to data to be assessed based on the risk evaluation model.
9. a kind of risk assessment equipment, including memory, processor and storage are in the memory and can be in the processing
The computer program run on device, which is characterized in that the processor realizes such as claim 1 when executing the computer program
The step of to methods of risk assessment described in any one of 7.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In the step of realization methods of risk assessment as described in any one of claims 1 to 7 when the computer program is executed by processor
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810414291.7A CN108876589A (en) | 2018-05-03 | 2018-05-03 | Methods of risk assessment, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810414291.7A CN108876589A (en) | 2018-05-03 | 2018-05-03 | Methods of risk assessment, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108876589A true CN108876589A (en) | 2018-11-23 |
Family
ID=64327347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810414291.7A Pending CN108876589A (en) | 2018-05-03 | 2018-05-03 | Methods of risk assessment, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108876589A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110059854A (en) * | 2019-03-13 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Method and device for risk identification |
| CN110136124A (en) * | 2019-05-17 | 2019-08-16 | 江门市中心医院 | A kind of pleura contact Lung neoplasm dividing method based on Robust Speed function |
-
2018
- 2018-05-03 CN CN201810414291.7A patent/CN108876589A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110059854A (en) * | 2019-03-13 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Method and device for risk identification |
| CN110136124A (en) * | 2019-05-17 | 2019-08-16 | 江门市中心医院 | A kind of pleura contact Lung neoplasm dividing method based on Robust Speed function |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pang et al. | An innovative neural network approach for stock market prediction | |
| Wang et al. | Grey wolf optimization evolving kernel extreme learning machine: Application to bankruptcy prediction | |
| WO2021078027A1 (en) | Method and apparatus for constructing network structure optimizer, and computer-readable storage medium | |
| CN109002845B (en) | Fine-grained image classification method based on deep convolutional neural network | |
| TWI726341B (en) | Sample attribute evaluation model training method, device, server and storage medium | |
| Kim et al. | Pan evaporation modeling using neural computing approach for different climatic zones | |
| WO2021169115A1 (en) | Risk control method, apparatus, electronic device, and computer-readable storage medium | |
| Kang et al. | A weight-incorporated similarity-based clustering ensemble method based on swarm intelligence | |
| Zhang et al. | The Bayesian additive classification tree applied to credit risk modelling | |
| CN112949837A (en) | Target recognition federal deep learning method based on trusted network | |
| CN110968701A (en) | Relationship map establishing method, device and equipment for graph neural network | |
| CN112966114B (en) | Literature classification method and device based on symmetrical graph convolutional neural network | |
| Chen et al. | Learning graph normalization for graph neural networks | |
| CN110532417A (en) | Image search method, device and terminal device based on depth Hash | |
| CN109766454A (en) | A kind of investor's classification method, device, equipment and medium | |
| CN102982373A (en) | OIN (Optimal Input Normalization) neural network training method for mixed SVM (Support Vector Machine) regression algorithm | |
| Zhang et al. | Forecasting of e-commerce transaction volume using a hybrid of extreme learning machine and improved moth-flame optimization algorithm | |
| CN109816177A (en) | A kind of Load aggregation quotient short-term load forecasting method, device and equipment | |
| CN109670927A (en) | The method of adjustment and its device of credit line, equipment, storage medium | |
| CN108256630A (en) | A kind of over-fitting solution based on low dimensional manifold regularization neural network | |
| Liang et al. | Survey of graph neural networks and applications | |
| CN115907001A (en) | Knowledge distillation-based federal diagram learning method and automatic driving method | |
| Kayhomayoon et al. | A new approach for regional groundwater level simulation: clustering, simulation, and optimization | |
| Wu et al. | Estimating fund-raising performance for start-up projects from a market graph perspective | |
| CN108876589A (en) | Methods of risk assessment, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181123 |
|
| RJ01 | Rejection of invention patent application after publication |