CN109583466A - Object ranking method, device, equipment and computer readable storage medium - Google Patents
Object ranking method, device, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN109583466A CN109583466A CN201811152500.1A CN201811152500A CN109583466A CN 109583466 A CN109583466 A CN 109583466A CN 201811152500 A CN201811152500 A CN 201811152500A CN 109583466 A CN109583466 A CN 109583466A
- Authority
- CN
- China
- Prior art keywords
- feature
- score
- exception
- probability
- class
- 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
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000002159 abnormal effect Effects 0.000 claims description 20
- 230000005856 abnormality Effects 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 239000000284 extract Substances 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 6
- 241000406668 Loxodonta cyclotis Species 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 14
- 230000007246 mechanism Effects 0.000 description 30
- 230000008569 process Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 12
- 230000006870 function Effects 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 8
- 230000006854 communication Effects 0.000 description 6
- 238000003064 k means clustering Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000001419 dependent effect Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 230000005484 gravity Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 230000002269 spontaneous effect Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000007717 exclusion Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000005086 pumping Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present disclosure provides object ranking method, device, equipment and computer readable storage medium.Object ranking method includes: the feature for extracting multiple objects as sample and acquiring the multiple object, wherein a part in the multiple object is marked as exception object and another part is marked as normal subjects；By presetting prediction model, predict that each object in the multiple object is the probability of exception object using the feature of the multiple object, and it is the probability of exception object according to each object in the multiple object predicted, calculates the score of each object；By presetting Clustering Model, the score based on each object clusters the multiple object；According to the score for the object for clustering resulting multiple classes, the grade of the object of each class is assessed, can be influenced to avoid artificial subjective factor, and dynamic grading effect can be reached, so that risk caused by the object being rated minimizes as far as possible.
Description
Technical field
The embodiment of the present disclosure is related to field of computer technology more particularly to object ranking method, device, equipment and computer
Readable storage medium storing program for executing.
Background technique
With the development of internet, more and more objects are had to be associated with external object, are occurred with external object
Associated degree of going deep into is depended on to the result of the reliability assessment of object.But in reliability estimation method mainly at present
Assign certain weight and score value to variation using expertise, the result of this method assessment be completely dependent on it is artificial it is subjective because
Element influences, and is unable to reach dynamic effect, can not capture in time to the variation of object.
Therefore, for facing for the object that association aspect integrity problem occurs, need one kind avoid it is artificial subjective because
Element influences, and can reach the object rating scheme of dynamic effect.
Summary of the invention
In view of this, disclosure first aspect provides a kind of object ranking method, comprising:
Multiple objects are extracted as sample and acquire the feature of the multiple object, wherein one in the multiple object
Part is marked as exception object and another part is marked as normal subjects；
By presetting prediction model, predict that each object in the multiple object is using the feature of the multiple object
The probability of exception object, and be the probability of exception object according to each object in the multiple object predicted, it calculates
The score of each object；
By presetting Clustering Model, the score based on each object clusters the multiple object；
According to the score for the object for clustering resulting multiple classes, the grade of the object of each class is assessed.
Disclosure second aspect provides a kind of object grading device, comprising:
Sampling and acquisition module, are configured as extracting multiple objects as sample and acquire the feature of the multiple object,
Wherein, a part in the multiple object is marked as exception object and another part is marked as normal subjects；
Predicting abnormality module is configured as described in the feature prediction by presetting prediction model, using the multiple object
Each object in multiple objects is the probability of exception object, and according to each object in the multiple object predicted
For the probability of exception object, the score of each object is calculated；
Cluster module is configured as by presetting Clustering Model, and the score based on each object is to the multiple right
As being clustered；
Grading module is configured as assessing the object of each class according to the score for the object for clustering resulting multiple classes
Grade.
The disclosure third aspect provides a kind of electronic equipment, including memory and processor；Wherein, the memory is used
In storing one or more computer instruction, wherein one or more computer instruction is executed by the processor with reality
Now method as described in relation to the first aspect.
Disclosure fourth aspect provides a kind of computer readable storage medium, is stored thereon with computer instruction, the meter
Method as described in relation to the first aspect is realized in the instruction of calculation machine when being executed by processor.
In disclosure embodiment, by extracting multiple objects as sample and acquiring the feature of the multiple object,
Wherein, a part in the multiple object is marked as exception object and another part is marked as normal subjects；
By presetting prediction model, using extracting multiple objects as sample and acquire the feature of the multiple object,
In, a part in the multiple object is marked as exception object and another part is marked as normal subjects；By pre-
If prediction model, predict that each object in the multiple object is the general of exception object using the feature of the multiple object
Rate, and be the probability of exception object according to each object in the multiple object predicted, calculate each object
Score；By presetting Clustering Model, the score based on each object clusters the multiple object；According to cluster
The score of the object of resulting multiple classes, assesses the grade of the object of each class, can predict that object is different with object-based feature
It normal probability and gives a mark to object, further according to the spontaneous cluster of scoring event, obtains object classification situation, and then comment object
Grade, avoids the process of all artificial judgments, so that artificial subjective factor be avoided to influence, and can reach dynamic grading effect, make
Risk caused by the object that must be rated minimizes as far as possible.
These aspects or other aspects of the disclosure can more straightforwards in the following description.
Detailed description of the invention
Technical solution in order to illustrate more clearly of the embodiment of the present disclosure or in the related technology, below will be to exemplary implementation
Attached drawing needed in example or description of Related Art is briefly described, it should be apparent that, the accompanying drawings in the following description
It is some exemplary embodiments of the disclosure, for those of ordinary skill in the art, before not making the creative labor
It puts, is also possible to obtain other drawings based on these drawings.
Fig. 1 shows the flow chart of the object ranking method according to one embodiment of the disclosure；
Fig. 2 shows the exemplary flow charts of the step S102 in the object ranking method according to one embodiment of the disclosure；
Fig. 3 shows the structural block diagram according to the object of one embodiment of disclosure grading device；
Fig. 4 is shown according to the exemplary of the predicting abnormality module 302 in the object of one embodiment of disclosure grading device
Structural block diagram；
Fig. 5 shows the structural block diagram of the equipment according to one embodiment of the disclosure；
Fig. 6 is adapted for the structure for realizing the computer system of the object ranking method according to one embodiment of the disclosure
Schematic diagram.
Specific embodiment
In order to make those skilled in the art more fully understand disclosure scheme, below in conjunction with the exemplary implementation of the disclosure
Attached drawing in example, is clearly and completely described the technical solution in disclosure exemplary embodiment.
In some processes of the description in the specification and claims of the disclosure and above-mentioned attached drawing, contain according to
Multiple operations that particular order occurs, but it should be clearly understood that these operations can not be what appears in this article suitable according to its
Sequence is executed or is executed parallel, and serial number of operation such as 101,102 etc. is only used for distinguishing each different operation, serial number
It itself does not represent and any executes sequence.In addition, these processes may include more or fewer operations, and these operations can
To execute or execute parallel in order.It should be noted that the description such as " first " herein, " second ", is for distinguishing not
Same message, equipment, module etc., does not represent sequencing, does not also limit " first " and " second " and be different type.
Below in conjunction with the attached drawing in disclosure exemplary embodiment, to the technical solution in disclosure exemplary embodiment
It being clearly and completely described, it is clear that described exemplary embodiment is only disclosure a part of the embodiment, rather than
Whole embodiments.Based on the embodiment in the disclosure, those skilled in the art institute without creative efforts
The every other embodiment obtained belongs to the range of disclosure protection..
Fig. 1 shows the flow chart of the object ranking method according to one embodiment of the disclosure.This method may include step
S101, S102, S103 and S104.
In step s101, multiple objects are extracted as sample and acquire the feature of multiple objects, wherein in multiple objects
A part be marked as exception object and another part is marked as normal subjects.
In step s 102, it by presetting prediction model, is predicted using the feature of multiple objects each in multiple objects
Object is the probability of exception object, and is the probability of exception object, meter according to each object in the multiple objects predicted
Calculate the score of each object.
In step s 103, by presetting Clustering Model, multiple objects are clustered based on the score of each object.
In step S104, according to the score for the object for clustering resulting multiple classes, the grade of the object of each class is assessed.
In one embodiment of the present disclosure, step S101 includes: according to labeled exception object relative to labeled
The modes that are in preset range of ratio of normal subjects extract multiple objects as sample and acquire the feature of multiple objects.
Extract multiple objects work that wherein exception object ratio meets preset condition (for example, exception object and normal subjects proportional balancing method)
For sample, more accurately the probability that object is exception object can be predicted.
In one embodiment of the present disclosure, the feature of object includes object property characteristics and object performance characteristic, wherein
The object property characteristics are to indicate the feature of the inherent attribute of the object, and object performance is characterized in described in expression
State from object to external presentation feature.
In one embodiment of the present disclosure, illustrate by taking the assessment to mechanism exception as an example according to disclosure embodiment
Scheme.In the related technology, certain weight is mainly assigned to variation using expertise in the rating scheme of mechanism
And score value, the rating result of the program is completely dependent on artificial subjective factor influence, and is unable to reach dynamic effect, to mechanism
Variation can not capture in time.In embodiment of the present disclosure, extract multiple mechanisms, some of them for had record of bad behavior and
There is the mechanism of larger impact, these mechanisms are labelled as abnormal mechanism, other mechanisms are normal mechanism, and Liang Zhong mechanism quantity is full
Sufficient preset condition, for example, keeping balance.Collecting mechanism essential information feature (i.e. object property characteristics) and cooperation performance characteristic
(i.e. object performance characteristic).For example, the essential information feature of mechanism includes such as number of users, headcount, sets up time, note
Whether volume capital is in debt.Performance characteristic during institution cooperation, for example, there is carriage several times with during other object cooperations
By crisis, whether number of users sharply glides in nearly one month, if has great organization change etc..
It is thus understood that object property characteristics are the objective attributive character of object, performance is object objective reality
Attribute.Object, which shows, is characterized by the feature of certain subjectivity, and performance is a degree of subjective assessment to object.
However, although object shows the existing a degree of subjective assessment to object of mark sheet, according to embodiment of the present disclosure
Technical solution by the subjective assessment extraction be characterized, input prediction model come predict object be exception object probability.Thus
It can be to avoid artificial subjective judgement, defect that is not objective enough and being unable to reach dynamic effect.
In embodiment of the present disclosure, using extracting multiple objects as sample and acquire the spy of the multiple object
Sign, wherein a part in the multiple object is marked as exception object and another part is marked as normal subjects；It is logical
Default prediction model is crossed, predicts that each object in the multiple object is exception object using the feature of the multiple object
Probability, and be the probability of exception object according to each object in the multiple object predicted, it is described each right to calculate
The score of elephant；By presetting Clustering Model, the score based on each object clusters the multiple object；According to poly-
The score of the object of the resulting multiple classes of class, assesses the grade of the object of each class, can predict object with object-based feature
It abnormal probability and gives a mark to object, further according to the spontaneous cluster of scoring event, obtains object classification situation, and then comment object
Grade, avoids the process of all artificial judgments, so that artificial subjective factor be avoided to influence, and can reach dynamic grading effect, make
Risk caused by the object that must be rated minimizes as far as possible.
Fig. 2 shows the exemplary flow charts of the step S102 in the object ranking method according to one embodiment of the disclosure.
As shown in Fig. 2, step S102 includes step S201 and S202.
In step s 201, by preset model, each object in multiple objects is predicted using the feature of multiple objects
For the probability of exception object, wherein the feature for the object being predicted is by the independent variable as preset model.
In step S202, according to default marking algorithm, it is abnormal for utilizing each object in the multiple objects predicted
The score of the probability calculation each object of object.
In one embodiment of the present disclosure, the label of multiple objects is rewritten as 0,1 variable.For example, 1 represent object as
Exception object, 0 to represent object be not exception object.The label of object is added in preset model as dependent variable.One
In a example, can there will be the feature of classified variable to change into 0,1 variable with the situation of one-hot coding in feature, as independent variable
It is added in model.
In the related technology, one-hot coding (One-Hot coding) is also known as an efficient coding, and method is using N shapes
State register encodes N number of state, and each state is by his independent register-bit, and when any, wherein
Only one effectively.For each feature, if it has m probable value, after one-hot coding, m have been reformed into
Binary feature.Also, these feature mutual exclusions, every time only one activation.
It can predict that each object in multiple objects is abnormal using the feature of multiple objects by Logic Regression Models
The probability of object.The example of Logic Regression Models such as following formula (1):
Wherein, p is the probability that object is exception object, that is, the probability that the label of object is；α is constant；X is object
Feature；β is the coefficient of individual features；The dimension that n is characterized, n are the positive integer greater than 1.
It will be understood by those skilled in the art that Logic Regression Models are predicted that object is exception object as preset model
Probability be only example, according to the introduction of the disclosure, can use various models as preset model to be different to object
The probability of normal object is predicted.
In one embodiment of the present disclosure, presetting marking algorithm can be following formula (2)
Score=100* (probability that 1- object is exception object) (2)
Wherein, Score is the score of object.
For example, if the probability that an object is exception object to be that 60%, 1-60% does not occur as abnormal general
Rate is 40%, by 100 be allocated as on the basis of point, 100*40%=40, the object score is 40 points.
It will be understood by those skilled in the art that can be utilized by various algorithms it is each right in the multiple objects predicted
As the score of the probability calculation each object for exception object.For example, being 60% for the probability that the above object is exception object
Example, can also be the probability 60% or 100* object of exception object by object using 1-60% as the score of object
For score of the probability 60% as object of exception object.As long as that is, can with the unified mode computing object score of standard with
It grades, object score can be calculated using various scoring models for object.
In one embodiment of the present disclosure, default Clustering Model can use various Clustering Models.For example, can use
K-means Clustering Model is as default Clustering Model.K-means algorithm is the process of the mobile class central point of a repetition, class
Central point, also referred to as center of gravity (centroids) is moved to the mean place it includes member, then repartition inside it at
Member.In one example, when being clustered using the score of object to multiple objects, can add using object score as independent variable
Enter in Clustering Model, and carry out normalization operation, i.e., its value size is transformed between 0,1, to be suitble to the measurement of distance.
Standardize formula such as following formula (3)
Y=(Score-min (Score))/(max (Score)-min (Score)) (3)
Wherein, y is the normalized value of equal part.
In the disclosure using k-means Clustering Model as in the embodiment of default Clustering Model, according to it is preset will be right
As number of classifying, k value is determined.Ancon rule can be used in this programme, i.e., draw the cost function of different value of K.With k value
Increase, average distortion degree can become smaller, it is every it is a kind of in number of samples can reduce therewith.But as k value increases, it may appear that
One apparent inflection point, the k that this point can be taken corresponding is as the classification number to distinguish to object.Wherein, each class
Distortion degree is equal to the quadratic sum of such center of gravity and its internal members' positional distance.If the member inside class the compact to each other
The distortion degree of class is smaller, conversely, the distortion degree of class is bigger if dispersing to each other by the member inside class.
When being clustered, can be related to obtaining the calculating of distance.In one embodiment of the present disclosure, it can use various
The distance between node when calculating cluster apart from calculation.For example, Europe can be used when carrying out k-means cluster
Family name's distance is carried out apart from calculating.
It will be understood by those skilled in the art that being only example, root using k-means Clustering Model as default Clustering Model
According to the introduction of the disclosure, various Clustering Models can be used as default Clustering Model to cluster to object score.
In one embodiment of the present disclosure, step S103 include: according to the score for the object for clustering resulting multiple classes,
The average equal part of the object of each class is calculated to assess the grade of the object of each class.
In one embodiment of the present disclosure, according to the score for the object for clustering resulting multiple classes, each class is calculated
The average equal part of object can be with to assess the hierarchical manner of the object of each class and arrange being averaged for every a kind of mechanism from big to small
Point, according to score value size divided rank to obtain object rating result.
It should be noted that different according to the calculation of the score of object, the object by calculating each class is averaged
Equal part can be different come the physical meaning of the grade of the object for each class assessed.
For example, calculated score value is higher when according to above formula (2) computing object score, then object gets over " normal ".Instead
It, if divided when using the probability that the probability or 100* object that object is exception object are exception object as when the score of object
Value is higher, then object gets over "abnormal".
It is appreciated that embodiment of the present disclosure and be not concerned with score height illustrate object grading height, but
The size for occurring abnormal probability by model automatic identification object is focused on, and then object is given a mark and graded accordingly, process
It is more objective.Moreover, embodiment of the present disclosure can also according to the data variation in the cooperation and associated process between object,
That is, the variation of object performance characteristic, predicts the probability size variation that object is abnormal, dynamic clustering, to reach dynamic in real time
Grading effect solves the problems, such as manually to grade not in time.
For example, the example assessed extremely according to the object evaluation method of disclosure embodiment mechanism in application
In, abnormal probability size can occur by presetting prediction model automatic identification mechanism, process is more objective.And it can root
According to the data variation during institution cooperation, the probability size variation that real-time projecting body is abnormal, dynamic clustering, to reach
Dynamic grading effect, solves the problems, such as artificial special mention agency ratings not in time.
Fig. 3 shows the structural block diagram according to the object of one embodiment of disclosure grading device.The apparatus may include pumpings
Sample and acquisition module 301, predicting abnormality module 302, cluster module 303 and grading module 304.
Sampling and acquisition module 301 are configured as extracting multiple objects as sample and acquire the feature of multiple objects,
In, a part in multiple objects is marked as exception object and another part is marked as normal subjects.
Predicting abnormality module 302 is configured as by presetting prediction model, and it is multiple right to be predicted using the feature of multiple objects
Each object as in is the probability of exception object, and is exception object according to each object in the multiple objects predicted
Probability, calculate the score of each object.
Cluster module 303 is configured as by presetting Clustering Model, is carried out based on the score of each object to multiple objects
Cluster.
Grading module 304 is configured as assessing the object of each class according to the score for the object for clustering resulting multiple classes
Grade.
In one embodiment of the present disclosure, sampling and acquisition module 301 are also configured to right according to labeled exception
As mode that the ratio relative to labeled normal subjects is in preset range extracts multiple objects as sample and acquires
The feature of multiple objects.It extracts wherein exception object ratio and meets preset condition (for example, exception object and normal subjects ratio are flat
Weighing apparatus) multiple objects be used as sample, more accurately object can be predicted for the probability of exception object.
In one embodiment of the present disclosure, the feature of object includes object property characteristics and object performance characteristic, wherein
The object property characteristics are to indicate the feature of the inherent attribute of the object, and object performance is characterized in described in expression
State from object to external presentation feature.
In one embodiment of the present disclosure, illustrate by taking the assessment to mechanism exception as an example according to disclosure embodiment
Scheme.In the related technology, certain weight is mainly assigned to variation using expertise in the rating scheme of mechanism
And score value, the rating result of the program is completely dependent on artificial subjective factor influence, and is unable to reach dynamic effect, to mechanism
Variation can not capture in time.In embodiment of the present disclosure, extract multiple mechanisms, some of them for had record of bad behavior and
There is the mechanism of larger impact, these mechanisms are labelled as abnormal mechanism, other mechanisms are normal mechanism, and Liang Zhong mechanism quantity is full
Sufficient preset condition, for example, keeping balance.Collecting mechanism essential information feature (i.e. object property characteristics) and cooperation performance characteristic
(i.e. object performance characteristic).For example, the essential information feature of mechanism includes such as number of users, headcount, sets up time, note
Whether volume capital is in debt.Performance characteristic during institution cooperation, for example, there is carriage several times with during other object cooperations
By crisis, whether number of users sharply glides in nearly one month, if has great organization change etc..
It is thus understood that object property characteristics are the objective attributive character of object, performance is object objective reality
Attribute.Object, which shows, is characterized by the feature of certain subjectivity, and performance is a degree of subjective assessment to object.
However, although object shows the existing a degree of subjective assessment to object of mark sheet, according to embodiment of the present disclosure
Technical solution by the subjective assessment extraction be characterized, input prediction model come predict object be exception object probability.Thus
It can be to avoid artificial subjective judgement, defect that is not objective enough and being unable to reach dynamic effect.
In embodiment of the present disclosure, using sampling and acquisition module, it is configured as extracting multiple objects as sample
And acquire the feature of the multiple object, wherein a part in the multiple object is marked as exception object and another
Part is marked as normal subjects；Predicting abnormality module is configured as utilizing the multiple object by presetting prediction model
Feature predicts that each object in the multiple object is the probability of exception object, and according to the multiple object predicted
In each object be exception object probability, calculate the score of each object；Cluster module is configured as by default
Clustering Model, the score based on each object cluster the multiple object；Grading module, is configured as according to poly-
The score of the object of the resulting multiple classes of class, assesses the grade of the object of each class, can predict object with object-based feature
It abnormal probability and gives a mark to object, further according to the spontaneous cluster of scoring event, obtains object classification situation, and then comment object
Grade, avoids the process of all artificial judgments, so that artificial subjective factor be avoided to influence, and can reach dynamic grading effect, make
Risk caused by the object that must be rated minimizes as far as possible.
Fig. 4 is shown according to the exemplary of the predicting abnormality module 302 in the object of one embodiment of disclosure grading device
Structural block diagram.As shown in figure 4, predicting abnormality module 302 includes predicting abnormality submodule 401 and marking submodule 402.
Predicting abnormality submodule 401 is configured as through preset model, predicts multiple objects using the feature of multiple objects
In each object be exception object probability, wherein the feature for the object being predicted is by the independent variable as preset model.
Marking submodule 402 is configured as according to default marking algorithm, and it is each right in the multiple objects predicted to utilize
As the score of the probability calculation each object for exception object.
In one embodiment of the present disclosure, the label of multiple objects is rewritten as 0,1 variable.For example, 1 represent object as
Exception object, 0 to represent object be not exception object.The label of object is added in preset model as dependent variable.One
In a example, can there will be the feature of classified variable to change into 0,1 variable with the situation of one-hot coding in feature, as independent variable
It is added in model.
In the related technology, one-hot coding (One-Hot coding) is also known as an efficient coding, and method is using N shapes
State register encodes N number of state, and each state is by his independent register-bit, and when any, wherein
Only one effectively.For each feature, if it has m probable value, after one-hot coding, m have been reformed into
Binary feature.Also, these feature mutual exclusions, every time only one activation.
It can predict that each object in multiple objects is abnormal using the feature of multiple objects by Logic Regression Models
The probability of object.The example of Logic Regression Models such as following formula (1):
Wherein, p is the probability that object is exception object, that is, the probability that the label of object is；α is constant；X is object
Feature；β is the coefficient of individual features；The dimension that n is characterized, n are the positive integer greater than 1.
It will be understood by those skilled in the art that Logic Regression Models are predicted that object is exception object as preset model
Probability be only example, according to the introduction of the disclosure, can use various models as preset model to be different to object
The probability of normal object is predicted.
In one embodiment of the present disclosure, presetting marking algorithm can be following formula (2)
Score=100* (probability that 1- object is exception object) (2)
Wherein, Score is the score of object.
For example, if the probability that an object is exception object to be that 60%, 1-60% does not occur as abnormal general
Rate is 40%, by 100 be allocated as on the basis of point, 100*40%=40, the object score is 40 points.
It will be understood by those skilled in the art that can be utilized by various algorithms it is each right in the multiple objects predicted
As the score of the probability calculation each object for exception object.For example, being 60% for the probability that the above object is exception object
Example, can also be the probability 60% or 100* object of exception object by object using 1-60% as the score of object
For score of the probability 60% as object of exception object.As long as that is, can with the unified mode computing object score of standard with
It grades, object score can be calculated using various scoring models for object.
In one embodiment of the present disclosure, default Clustering Model can use various Clustering Models.For example, can use
K-means Clustering Model is as default Clustering Model.K-means algorithm is the process of the mobile class central point of a repetition, class
Central point, also referred to as center of gravity (centroids) is moved to the mean place it includes member, then repartition inside it at
Member.In one example, when being clustered using the score of object to multiple objects, can add using object score as independent variable
Enter in Clustering Model, and carry out normalization operation, i.e., its value size is transformed between 0,1, to be suitble to the measurement of distance.
Standardize formula such as following formula (3)
Y=(Score-min (Score))/(max (Score)-min (Score)) (3)
Wherein, y is the normalized value of equal part.
In the disclosure using k-means Clustering Model as in the embodiment of default Clustering Model, according to it is preset will be right
As number of classifying, k value is determined.Ancon rule can be used in this programme, i.e., draw the cost function of different value of K.With k value
Increase, average distortion degree can become smaller, it is every it is a kind of in number of samples can reduce therewith.But as k value increases, it may appear that
One apparent inflection point, the k that this point can be taken corresponding is as the classification number to distinguish to object.Wherein, each class
Distortion degree is equal to the quadratic sum of such center of gravity and its internal members' positional distance.If the member inside class the compact to each other
The distortion degree of class is smaller, conversely, the distortion degree of class is bigger if dispersing to each other by the member inside class.
When being clustered, can be related to obtaining the calculating of distance.In one embodiment of the present disclosure, it can use various
The distance between node when calculating cluster apart from calculation.For example, Europe can be used when carrying out k-means cluster
Family name's distance is carried out apart from calculating.
It will be understood by those skilled in the art that being only example, root using k-means Clustering Model as default Clustering Model
According to the introduction of the disclosure, various Clustering Models can be used as default Clustering Model to cluster to object score.
In one embodiment of the present disclosure, step S103 include: according to the score for the object for clustering resulting multiple classes,
The average equal part of the object of each class is calculated to assess the grade of the object of each class.
In one embodiment of the present disclosure, according to the score for the object for clustering resulting multiple classes, each class is calculated
The average equal part of object can be with to assess the hierarchical manner of the object of each class and arrange being averaged for every a kind of mechanism from big to small
Point, according to score value size divided rank to obtain object rating result.
It should be noted that different according to the calculation of the score of object, the object by calculating each class is averaged
Equal part can be different come the physical meaning of the grade of the object for each class assessed.
For example, calculated score value is higher when according to above formula (2) computing object score, then object gets over " normal ".Instead
It, if divided when using the probability that the probability or 100* object that object is exception object are exception object as when the score of object
Value is higher, then object gets over "abnormal".
It is appreciated that embodiment of the present disclosure and be not concerned with score height illustrate object grading height, but
The size for occurring abnormal probability by model automatic identification object is focused on, and then object is given a mark and graded accordingly, process
It is more objective.Moreover, embodiment of the present disclosure can also according to the data variation in the cooperation and associated process between object,
That is, the variation of object performance characteristic, predicts the probability size variation that object is abnormal, dynamic clustering, to reach dynamic in real time
Grading effect solves the problems, such as manually to grade not in time.
For example, the example assessed extremely according to the object evaluation method of disclosure embodiment mechanism in application
In, abnormal probability size can occur by presetting prediction model automatic identification mechanism, process is more objective.And it can root
According to the data variation during institution cooperation, the probability size variation that real-time projecting body is abnormal, dynamic clustering, to reach
Dynamic grading effect, solves the problems, such as artificial special mention agency ratings not in time.
The foregoing describe the built-in functions and structure of object grading device, in a possible design, object grading
The structure of device can realize for object grade equipment, as shown in Figure 5, the processing equipment 500 may include processor 501 and
Memory 502.
The memory 502 executes object ranking method in any of the above-described embodiment for storing support target grading device
Program, the processor 501 is configurable for executing the program stored in the memory 502.
The memory 502 is for storing one or more computer instruction, wherein one or more computer refers to
Order is executed by the processor 501.
The processor 501 is used to execute all or part of the steps in aforementioned approaches method step.
It wherein, can also include communication interface in the structure of the object grading equipment, for object grading equipment and its
His equipment or communication.
Disclosure exemplary embodiment additionally provides a kind of computer storage medium, for storing the object grading device
Computer software instructions used, it includes for executing program involved in object ranking method in any of the above-described embodiment.
Fig. 6 is adapted for the structure for realizing the computer system of the object ranking method according to one embodiment of the disclosure
Schematic diagram.
As shown in fig. 6, computer system 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in
Program in memory (ROM) 602 or be loaded into the program in random access storage device (RAM) 603 from storage section 608 and
Execute the various processing in above-mentioned embodiment shown in FIG. 1.In RAM603, be also stored with system 600 operate it is required each
Kind program and data.CPU601, ROM602 and RAM603 are connected with each other by bus 604.Input/output (I/O) interface 605
It is also connected to bus 604.
I/O interface 605 is connected to lower component: the importation 606 including keyboard, mouse etc.；It is penetrated including such as cathode
The output par, c 607 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section 608 including hard disk etc.；
And the communications portion 609 of the network interface card including LAN card, modem etc..Communications portion 609 via such as because
The network of spy's net executes communication process.Driver 610 is also connected to I/O interface 605 as needed.Detachable media 611, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 610, in order to read from thereon
Computer program be mounted into storage section 608 as needed.
Particularly, according to embodiment of the present disclosure, it is soft to may be implemented as computer above with reference to Fig. 1 method described
Part program.For example, embodiment of the present disclosure includes a kind of computer program product comprising be tangibly embodied in and its readable
Computer program on medium, the computer program include the program code for executing the data processing method of Fig. 1.At this
In the embodiment of sample, which can be downloaded and installed from network by communications portion 609, and/or from can
Medium 611 is dismantled to be mounted.
Flow chart and block diagram in attached drawing illustrate system, method and computer according to the various embodiments of the disclosure
The architecture, function and operation in the cards of program product.In this regard, each box in course diagram or block diagram can be with
A part of a module, section or code is represented, a part of the module, section or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, and/or specialized hardware and meter can be used
The combination of calculation machine instruction is realized.
Being described in unit or module involved in disclosure embodiment can be realized by way of software, can also
It is realized in a manner of through hardware.Described unit or module also can be set in the processor, these units or module
Title do not constitute the restriction to the unit or module itself under certain conditions.
As on the other hand, the disclosure additionally provides a kind of computer readable storage medium, the computer-readable storage medium
Matter can be computer readable storage medium included in device described in above embodiment；It is also possible to individualism,
Without the computer readable storage medium in supplying equipment.Computer-readable recording medium storage has one or more than one journey
Sequence, described program is used to execute by one or more than one processor is described in disclosed method.
Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that invention scope involved in the disclosure, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic
Scheme, while should also cover in the case where not departing from the inventive concept, it is carried out by above-mentioned technical characteristic or its equivalent feature
Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed in the disclosure
Can technical characteristic replaced mutually and the technical solution that is formed.
Claims (12)
1. a kind of object ranking method characterized by comprising
Multiple objects are extracted as sample and acquire the feature of the multiple object, wherein a part in the multiple object
It is marked as exception object and another part is marked as normal subjects；
By presetting prediction model, predict that each object in the multiple object is abnormal using the feature of the multiple object
The probability of object, and be the probability of exception object according to each object in the multiple object predicted, described in calculating
The score of each object；
By presetting Clustering Model, the score based on each object clusters the multiple object；
According to the score for the object for clustering resulting multiple classes, the grade of the object of each class is assessed.
2. the method according to claim 1, wherein described extract multiple objects as sample and acquire described more
The feature of a object, comprising:
It is extracted in such a way that labeled exception object is in preset range relative to the ratio of labeled normal subjects
Multiple objects are as sample and acquire the feature of the multiple object.
3. the method according to claim 1, wherein the feature of the object includes object property characteristics and object
Performance characteristic, wherein the object property characteristics are to indicate the feature of the inherent attribute of the object, and the object shows
Be characterized in indicating state from the object to external presentation feature.
4. the method according to claim 1, wherein described by presetting prediction model, using the multiple right
The feature of elephant predicts that each object in the multiple object is the probability of exception object, and the multiple according to what is predicted
Each object in object is the probability of exception object, calculates the score of each object, comprising:
By preset model, predict that each object in the multiple object is exception object using the feature of the multiple object
Probability, wherein the feature for the object being predicted is by the independent variable as the preset model；
According to default marking algorithm, utilizing each object in the multiple object predicted is the probability calculation of exception object
The score of each object.
5. the method according to claim 1, wherein the obtaining according to the object for clustering resulting multiple classes
Point, assess the grade of the object of each class, comprising:
According to the score for the object for clustering resulting multiple classes, the average equal part of the object of each class is calculated to assess each class
The grade of object.
The device 6. a kind of object is graded characterized by comprising
Sampling and acquisition module, are configured as extracting multiple objects as sample and acquire the feature of the multiple object, wherein
A part in the multiple object is marked as exception object and another part is marked as normal subjects；
Predicting abnormality module is configured as predicting using the feature of the multiple object the multiple by presetting prediction model
Each object in object is the probability of exception object, and is different according to each object in the multiple object predicted
The probability of normal object, calculates the score of each object；
Cluster module, be configured as by preset Clustering Model, the score based on each object to the multiple object into
Row cluster；
Grading module is configured as assessing the grade of the object of each class according to the score for the object for clustering resulting multiple classes.
7. device according to claim 6, which is characterized in that the sampling and acquisition module are also configured to
It is extracted in such a way that labeled exception object is in preset range relative to the ratio of labeled normal subjects
Multiple objects are as sample and acquire the feature of the multiple object.
8. device according to claim 6, which is characterized in that the feature of the object includes object property characteristics and object
Performance characteristic, wherein the object property characteristics are to indicate the feature of the inherent attribute of the object, and the object shows
Be characterized in indicating state from the object to external presentation feature.
9. device according to claim 6, which is characterized in that the predicting abnormality module includes:
Predicting abnormality submodule is configured as through preset model, and it is the multiple right to be predicted using the feature of the multiple object
Each object as in is the probability of exception object, wherein the feature for the object being predicted is by becoming certainly as the preset model
Amount；
Marking submodule is configured as being predicted using the predicting abnormality submodule described more according to default marking algorithm
Each object in a object is the score of each object described in the probability calculation of exception object.
10. device according to claim 6, which is characterized in that the grading module is also configured to
The score that the object of resulting multiple classes is clustered according to the cluster module, calculate the average equal part of the object of each class with
Assess the grade of the object of each class.
11. a kind of electronic equipment, which is characterized in that including memory and processor；Wherein, the memory is for storing one
Or a plurality of computer instruction, wherein one or more computer instruction is executed by the processor to realize that right such as is wanted
Seek the described in any item methods of 1-5.
12. a kind of computer readable storage medium, is stored thereon with computer instruction, which is characterized in that the computer instruction quilt
Processor realizes the method according to claim 1 to 5 when executing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811152500.1A CN109583466A (en) | 2018-09-29 | 2018-09-29 | Object ranking method, device, equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811152500.1A CN109583466A (en) | 2018-09-29 | 2018-09-29 | Object ranking method, device, equipment and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109583466A true CN109583466A (en) | 2019-04-05 |
Family
ID=65920232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811152500.1A Pending CN109583466A (en) | 2018-09-29 | 2018-09-29 | Object ranking method, device, equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109583466A (en) |
-
2018
- 2018-09-29 CN CN201811152500.1A patent/CN109583466A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110837931B (en) | Customer churn prediction method, device and storage medium | |
| CN108090800B (en) | Game prop pushing method and device based on player consumption potential | |
| CN105488539B (en) | The predictor method and device of the generation method and device of disaggregated model, power system capacity | |
| CN106651057A (en) | Mobile terminal user age prediction method based on installation package sequence table | |
| CN107622326A (en) | User's classification, available resources Forecasting Methodology, device and equipment | |
| KR20180097895A (en) | APPARATUS AND METHOD for DETECTING DATA OUTLIERS IN FRAUN DETECTION SYSTEM | |
| CN111611496A (en) | Product recommendation method and device | |
| CN115081025A (en) | Sensitive data management method and device based on digital middlebox and electronic equipment | |
| CN110708285A (en) | Flow monitoring method, device, medium and electronic equipment | |
| CN114610865A (en) | Method, device and equipment for recommending recalled text and storage medium | |
| CN113313538A (en) | User consumption capacity prediction method and device, electronic equipment and storage medium | |
| CN109345201A (en) | Human Resources Management Method, device, electronic equipment and storage medium | |
| US9594757B2 (en) | Document management system, document management method, and document management program | |
| CN112348685A (en) | Credit scoring method, device, equipment and storage medium | |
| CN113704389A (en) | Data evaluation method and device, computer equipment and storage medium | |
| CN112163962A (en) | Method and device for model training and business wind control | |
| CN115408527B (en) | Text classification method and device, electronic equipment and storage medium | |
| JP5427119B2 (en) | Similar document search device, similar document search method, program thereof, and recording medium | |
| CN109583466A (en) | Object ranking method, device, equipment and computer readable storage medium | |
| CN113435746B (en) | User workload scoring method and device, electronic equipment and storage medium | |
| CN110008974A (en) | Behavioral data prediction technique, device, electronic equipment and computer storage medium | |
| CN112258095B (en) | Standard normal distribution based scoring method, device, equipment and storage medium | |
| CN108985811A (en) | Method, apparatus and electronic equipment for precision marketing | |
| CN114139657B (en) | Guest group portrait generation method and device, electronic equipment and storage medium | |
| CN113688854A (en) | Data processing method and device and computing equipment |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200922 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. Effective date of registration: 20200922 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190405 |